Search

Many complex physical systems exhibit a rich variety of discrete behavioural modes. Often, the system complexity limits the applicability of standard modelling tools. Hence, understanding the underlying physics of different behaviours and distinguishing between them is challenging. Although traditional machine learning techniques could predict and classify behaviour well, typically they do not provide any meaningful insight into the underlying physics of the system. In this paper we present a novel method for extracting physically meaningful clusters of discrete behaviour from limited experimental observations. This method obtains a set of physically plausible functions that both facilitate behavioural clustering and aid in system understanding. We demonstrate the approach on the V-shaped falling paper system, a new falling paper type system that exhibits four distinct behavioural modes depending on a few morphological parameters. Using just 49 experimental observations, the method discovered a set of candidate functions that distinguish behaviours with an error of 2.04%, while also aiding insight into the physical phenomena driving each behaviour. [ABSTRACT FROM AUTHOR]

Fundamental scientific advances can take decades to translate into improvements in human health. Shortening this interval would increase the rate at which scientific discoveries lead to successful treatment of human disease. One way to accomplish this would be to identify which advances in knowledge are most likely to translate into clinical research. Toward that end, we built a machine learning system that detects whether a paper is likely to be cited by a future clinical trial or guideline. Despite the noisiness of citation dynamics, as little as 2 years of postpublication data yield accurate predictions about a paper's eventual citation by a clinical article (accuracy = 84%, F1 score = 0.56; compared to 19% accuracy by chance). We found that distinct knowledge flow trajectories are linked to papers that either succeed or fail to influence clinical research. Translational progress in biomedicine can therefore be assessed and predicted in real time based on information conveyed by the scientific community's early reaction to a paper. [ABSTRACT FROM AUTHOR]

This paper, based on differential privacy protecting K-means clustering algorithm, realizes privacy protection by adding data-disturbing Laplace noise to cluster center point. In order to solve the problem of Laplace noise randomness which causes the center point to deviate, especially when poor availability of clustering results appears because of small privacy budget parameters, an improved differential privacy protecting K-means clustering algorithm was raised in this paper. The improved algorithm uses the contour coefficients to quantitatively evaluate the clustering effect of each iteration and add different noise to different clusters. In order to be adapted to the huge number of data, this paper provides an algorithm design in MapReduce Framework. Experimental finding shows that the new algorithm improves the availability of the algorithm clustering results under the condition of ensuring individual privacy without significantly increasing its operating time. [ABSTRACT FROM AUTHOR]

This paper proposes SAR imaging algorithm with largest coherence based on the existing SAR imaging algorithm. The basic idea of SAR imaging algorithm in imaging processing is that output signal can have maximum signal-to-noise ratio (SNR) by using the optimal imaging parameters. Traditional imaging algorithm can acquire the best focusing effect, but would bring the decoherence phenomenon in subsequent interference process. Algorithm proposed in this paper is that SAR echo adopts consistent imaging parameters in focusing processing. Although the SNR of the output signal is reduced slightly, their coherence is ensured greatly, and finally the interferogram with high quality is obtained. In this paper, two scenes of Envisat ASAR data in Zhangbei are employed to conduct experiment for this algorithm. Compared with the interferogram from the traditional algorithm, the results show that this algorithm is more suitable for SAR interferometry (InSAR) research and application. [ABSTRACT FROM AUTHOR]

Increasingly, scholarly articles contain URI references to “web at large” resources including project web sites, scholarly wikis, ontologies, online debates, presentations, blogs, and videos. Authors reference such resources to provide essential context for the research they report on. A reader who visits a web at large resource by following a URI reference in an article, some time after its publication, is led to believe that the resource’s content is representative of what the author originally referenced. However, due to the dynamic nature of the web, that may very well not be the case. We reuse a dataset from a previous study in which several authors of this paper were involved, and investigate to what extent the textual content of web at large resources referenced in a vast collection of Science, Technology, and Medicine (STM) articles published between 1997 and 2012 has remained stable since the publication of the referencing article. We do so in a two-step approach that relies on various well-established similarity measures to compare textual content. In a first step, we use 19 web archives to find snapshots of referenced web at large resources that have textual content that is representative of the state of the resource around the time of publication of the referencing paper. We find that representative snapshots exist for about 30% of all URI references. In a second step, we compare the textual content of representative snapshots with that of their live web counterparts. We find that for over 75% of references the content has drifted away from what it was when referenced. These results raise significant concerns regarding the long term integrity of the web-based scholarly record and call for the deployment of techniques to combat these problems. [ABSTRACT FROM AUTHOR]

Biometric recognition technology based on eye-movement dynamics has been in development for more than ten years. Different visual tasks, feature extraction and feature recognition methods are proposed to improve the performance of eye movement biometric system. However, the correct identification and verification rates, especially in long-term experiments, as well as the effects of visual tasks and eye trackers’ temporal and spatial resolution are still the foremost considerations in eye movement biometrics. With a focus on these issues, we proposed a new visual searching task for eye movement data collection and a new class of eye movement features for biometric recognition. In order to demonstrate the improvement of this visual searching task being used in eye movement biometrics, three other eye movement feature extraction methods were also tested on our eye movement datasets. Compared with the original results, all three methods yielded better results as expected. In addition, the biometric performance of these four feature extraction methods was also compared using the equal error rate (EER) and Rank-1 identification rate (Rank-1 IR), and the texture features introduced in this paper were ultimately shown to offer some advantages with regard to long-term stability and robustness over time and spatial precision. Finally, the results of different combinations of these methods with a score-level fusion method indicated that multi-biometric methods perform better in most cases. [ABSTRACT FROM AUTHOR]

Recommender systems are vulnerable to shilling attacks. Forged user-generated content data, such as user ratings and reviews, are used by attackers to manipulate recommendation rankings. Shilling attack detection in recommender systems is of great significance to maintain the fairness and sustainability of recommender systems. The current studies have problems in terms of the poor universality of algorithms, difficulty in selection of user profile attributes, and lack of an optimization mechanism. In this paper, a shilling behaviour detection structure based on abnormal group user findings and rating time series analysis is proposed. This paper adds to the current understanding in the field by studying the credibility evaluation model in-depth based on the rating prediction model to derive proximity-based predictions. A method for detecting suspicious ratings based on suspicious time windows and target item analysis is proposed. Suspicious rating time segments are determined by constructing a time series, and data streams of the rating items are examined and suspicious rating segments are checked. To analyse features of shilling attacks by a group user’s credibility, an abnormal group user discovery method based on time series and time window is proposed. Standard testing datasets are used to verify the effect of the proposed method. [ABSTRACT FROM AUTHOR]

This paper addresses patient-reported outcomes (PROs) and telemonitoring in congestive heart failure (CHF), both increasingly important topics. The interest in CHF trials is shifting from hard end-points such as hospitalization and mortality, to softer end-points such health-related quality of life. However, the relation of these softer end-points to objective parameters is not well studied. Telemonitoring is suitable for collecting both patient-reported outcomes and objective parameters. Most telemonitoring studies, however, do not take full advantage of the available sensor technology and intelligent data analysis. The Chiron clinical observational study was performed among 24 CHF patients (17 men and 7 women, age 62.9 ± 9.4 years, 15 NYHA class II and 9 class III, 10 of ishaemic, aetiology, 6 dilated, 2 valvular, and 6 of multiple aetiologies or cardiomyopathy) in Italy and UK. A large number of physiological and ambient parameters were collected by wearable and other devices, together with PROs describing how well the patients felt, over 1,086 days of observation. The resulting data were mined for relations between the objective parameters and the PROs. The objective parameters (humidity, ambient temperature, blood pressure, SpO2, and sweeting intensity) could predict the PROs with accuracies up to 86% and AUC up to 0.83, making this the first report providing evidence for ambient and physiological parameters to be objectively related to PROs in CHF patients. We also analyzed the relations in the predictive models, gaining some insights into what affects the feeling of health, which was also generally not attempted in previous investigations. The paper strongly points to the possibility of using PROs as primary end-points in future trials. [ABSTRACT FROM AUTHOR]

Heart rate variability (HRV) analysis has quantified the functioning of the autonomic regulation of the heart and heart's ability to respond. However, majority of studies on HRV report several differences between patients with congestive heart failure (CHF) and healthy subjects, such as time-domain, frequency domain and nonlinear HRV measures. In the paper, we mainly presented a new approach to detect congestive heart failure (CHF) based on combination support vector machine (SVM) and three nonstandard heart rate variability (HRV) measures (e.g. SUM_TD, SUM_FD and SUM_IE). The CHF classification model was presented by using SVM classifier with the combination SUM_TD and SUM_FD. In the analysis performed, we found that the CHF classification algorithm could obtain the best performance with the CHF classification accuracy, sensitivity and specificity of 100%, 100%, 100%, respectively. [ABSTRACT FROM AUTHOR]

B-spline functions are widely used in many industrial applications such as computer graphic representations, computer aided design, computer aided manufacturing, computer numerical control, etc. Recently, there exist some demands, e.g. in reverse engineering (RE) area, to employ B-spline curves for non-trivial cases that include curves with discontinuous points, cusps or turning points from the sampled data. The most challenging task in these cases is in the identification of the number of knots and their respective locations in non-uniform space in the most efficient computational cost. This paper presents a new strategy for fitting any forms of curve by B-spline functions via local algorithm. A new two-step method for fast knot calculation is proposed. In the first step, the data is split using a bisecting method with predetermined allowable error to obtain coarse knots. Secondly, the knots are optimized, for both locations and continuity levels, by employing a non-linear least squares technique. The B-spline function is, therefore, obtained by solving the ordinary least squares problem. The performance of the proposed method is validated by using various numerical experimental data, with and without simulated noise, which were generated by a B-spline function and deterministic parametric functions. This paper also discusses the benchmarking of the proposed method to the existing methods in literature. The proposed method is shown to be able to reconstruct B-spline functions from sampled data within acceptable tolerance. It is also shown that, the proposed method can be applied for fitting any types of curves ranging from smooth ones to discontinuous ones. In addition, the method does not require excessive computational cost, which allows it to be used in automatic reverse engineering applications. [ABSTRACT FROM AUTHOR]

ICD-10(International Classification of Diseases 10th revision) is a classification of a disease, symptom, procedure, or injury. Diseases are often described in patients’ medical records with free texts, such as terms, phrases and paraphrases, which differ significantly from those used in ICD-10 classification. This paper presents an improved approach based on the Longest Common Subsequence (LCS) and semantic similarity for automatic Chinese diagnoses, mapping from the disease names given by clinician to the disease names in ICD-10. LCS refers to the longest string that is a subsequence of every member of a given set of strings. The proposed method of improved LCS in this paper can increase the accuracy of processing in Chinese disease mapping. [ABSTRACT FROM AUTHOR]

The process of Nuclei detection in high-grade breast cancer images is quite challenging in the case of image processing techniques due to certain heterogeneous characteristics of cancer nuclei such as enlarged and irregularly shaped nuclei, highly coarse chromatin marginalized to the nuclei periphery and visible nucleoli. Recent reviews state that existing techniques show appreciable segmentation accuracy on breast histopathology images whose nuclei are dispersed and regular in texture and shape; however, typical cancer nuclei are often clustered and have irregular texture and shape properties. This paper proposes a novel segmentation algorithm for detecting individual nuclei from Hematoxylin and Eosin (H&E) stained breast histopathology images. This detection framework estimates a nuclei saliency map using tensor voting followed by boundary extraction of the nuclei on the saliency map using a Loopy Back Propagation (LBP) algorithm on a Markov Random Field (MRF). The method was tested on both whole-slide images and frames of breast cancer histopathology images. Experimental results demonstrate high segmentation performance with efficient precision, recall and dice-coefficient rates, upon testing high-grade breast cancer images containing several thousand nuclei. In addition to the optimal performance on the highly complex images presented in this paper, this method also gave appreciable results in comparison with two recently published methods—Wienert et al. (2012) and Veta et al. (2013), which were tested using their own datasets. [ABSTRACT FROM AUTHOR]

This paper proposes a method to predict fluctuations in the prices of cryptocurrencies, which are increasingly used for online transactions worldwide. Little research has been conducted on predicting fluctuations in the price and number of transactions of a variety of cryptocurrencies. Moreover, the few methods proposed to predict fluctuation in currency prices are inefficient because they fail to take into account the differences in attributes between real currencies and cryptocurrencies. This paper analyzes user comments in online cryptocurrency communities to predict fluctuations in the prices of cryptocurrencies and the number of transactions. By focusing on three cryptocurrencies, each with a large market size and user base, this paper attempts to predict such fluctuations by using a simple and efficient method. [ABSTRACT FROM AUTHOR]

Traditional methods for monitoring influenza are haphazard and lack fine-grained details regarding the spatial and temporal dynamics of outbreaks. Twitter gives researchers and public health officials an opportunity to examine the spread of influenza in real-time and at multiple geographical scales. In this paper, we introduce an improved framework for monitoring influenza outbreaks using the social media platform Twitter. Relying upon techniques from geographic information science (GIS) and data mining, Twitter messages were collected, filtered, and analyzed for the thirty most populated cities in the United States during the 2013–2014 flu season. The results of this procedure are compared with national, regional, and local flu outbreak reports, revealing a statistically significant correlation between the two data sources. The main contribution of this paper is to introduce a comprehensive data mining process that enhances previous attempts to accurately identify tweets related to influenza. Additionally, geographical information systems allow us to target, filter, and normalize Twitter messages. [ABSTRACT FROM AUTHOR]

Background and objectives: ECG signal is relatively weak and vulnerable to various noise interferences, such as electromyography. There will be robustness problems when detecting the instantaneous heart rate independently. In some cases, multiple human physiologic parameters are monitored to help in heart rate detection. Methods: In this paper, an algorithm that marks the R-wave peaks with the help of simultaneously recorded continuous blood pressure is proposed and tested on two databases. One database, called the challenge database, is provided by the PhysioNet/Computing in Cardiology Challenge 2014, and the other is the MGH/MF waveform database. Results: The final scores of the proposed algorithm are 97.3% for the challenge database and 96.6% for the MGH/MF waveform database. Conclusions: The experimental results show that this algorithm has high detection accuracy and a relatively strong robustness. [ABSTRACT FROM AUTHOR]

This paper proposes a mixture regression model-based method for drug sensitivity prediction. The proposed method explicitly addresses two fundamental issues in drug sensitivity prediction, namely, population heterogeneity and feature selection pertaining to each of the subpopulations. The mixture regression model is estimated using the imputation-conditional consistency algorithm, and the resulting estimator is consistent. This paper also proposes an average-BIC criterion for determining the number of components for the mixture regression model. The proposed method is applied to the CCLE dataset, and the numerical results indicate that the proposed method can make a drastic improvement over the existing ones, such as random forest, support vector regression, and regularized linear regression, in both drug sensitivity prediction and feature selection. The p-values for the comparisons in drug sensitivity prediction can reach the order O(10−8) or lower for the drugs with heterogeneous populations. [ABSTRACT FROM AUTHOR]

A recurrent criticism concerning the use of online social media data in political science research is the lack of demographic information about social media users. By employing a face-recognition algorithm to the profile pictures of Facebook users, the paper derives two fundamental demographic characteristics (age and gender) of a sample of Facebook users who interacted with the most relevant British parties in the two weeks before the Brexit referendum of 23 June 2016. The article achieves the goals of (i) testing the precision of the algorithm, (ii) testing its validity, (iii) inferring new evidence on digital mobilisation, and (iv) tracing the path for future developments and application of the algorithm. The findings show that the algorithm is reliable and that it can be fruitfully used in political and social sciences both to confirm the validity of survey data and to obtain information from populations that are generally unavailable within traditional surveys. [ABSTRACT FROM AUTHOR]

Dementia is a neurological and cognitive condition that affects millions of people around the world. At any given time in the United Kingdom, 1 in 4 hospital beds are occupied by a person with dementia, while about 22% of these hospital admissions are due to preventable causes. In this paper we discuss using Internet of Things (IoT) technologies and in-home sensory devices in combination with machine learning techniques to monitor health and well-being of people with dementia. This will allow us to provide more effective and preventative care and reduce preventable hospital admissions. One of the unique aspects of this work is combining environmental data with physiological data collected via low cost in-home sensory devices to extract actionable information regarding the health and well-being of people with dementia in their own home environment. We have worked with clinicians to design our machine learning algorithms where we focused on developing solutions for real-world settings. In our solutions, we avoid generating too many alerts/alarms to prevent increasing the monitoring and support workload. We have designed an algorithm to detect Urinary Tract Infections (UTI) which is one of the top five reasons of hospital admissions for people with dementia (around 9% of hospital admissions for people with dementia in the UK). To develop the UTI detection algorithm, we have used a Non-negative Matrix Factorisation (NMF) technique to extract latent factors from raw observation and use them for clustering and identifying the possible UTI cases. In addition, we have designed an algorithm for detecting changes in activity patterns to identify early symptoms of cognitive decline or health decline in order to provide personalised and preventative care services. For this purpose, we have used an Isolation Forest (iForest) technique to create a holistic view of the daily activity patterns. This paper describes the algorithms and discusses the evaluation of the work using a large set of real-world data collected from a trial with people with dementia and their caregivers. [ABSTRACT FROM AUTHOR]

Research: The body composition model is closely related to the physiological characteristics of the human body. At the same time there can be a large number of physiological characteristics, many of which may be redundant or irrelevant. In existing human physiological feature selection algorithms, it is difficult to overcome the impact that redundancy and irrelevancy may have on human body composition modeling. This suggests a role for selection algorithms, where human physiological characteristics are identified using a combination of filtering and improved clustering. To do this, a feature filtering method based on Hilbert-Schmidt dependency criteria is first of all used to eliminate irrelevant features. After this, it is possible to use improved Chameleon clustering to increase the combination of sub-clusters amongst the characteristics, thereby removing any redundant features to obtain a candidate feature set for human body composition modeling. Method We report here on the use of an algorithm to filter the characteristic parameters in INBODY770 (this paper used INBODY 770 as body composition analyzer.) measurement data, which has three commonly-used impedance bands (1 kHZ, 250 kHZ, 500 kHZ). This algorithm is able to filter out parameters that have a low correlation with body composition BFM. The algorithm is also able to draw upon improved clustering techniques to reduce the initial feature set from 29 parameters to 10 parameters for any parameters of the 250 kHZ band that remain after filtering. In addition, we also examined the impact of different sample sizes on feature selection. Result: The proposed algorithm is able to remove irrelevant and redundant features and the resulting correlation between the model and the body composition (BFM which is a whole body fat evaluation can better assess the body's overall fat and muscle composition.) is 0.978, thereby providing an improved model for prediction with a relative error of less than 0.12. [ABSTRACT FROM AUTHOR]

In clinical neuropsychology the cognitive abilities of neurological patients are commonly estimated using well-established paper-based tests. Typically, scores on some tests remain relatively well preserved, whilst others exhibit a significant and disproportionate decline. Scores on those tests that measure preserved cognitive functions (so-called ‘hold’ tests) may be used to estimate premorbid abilities, including scores in non-hold tests that would have been expected prior to the onset of cognitive impairment. Many hold tests entail word reading, with each word being graded as correctly or incorrectly pronounced. Inevitably, such tests are likely to contain words that provide little or no diagnostic power (i.e., can be eliminated without negatively affecting prediction accuracy). In this paper, a genetic algorithm is developed and demonstrated, using n = 92 neurologically healthy participants, to identify optimal word subsets from the National Adult Reading Test that minimize the mean error in predicting the most widely used clinical measure of IQ and cognitive ability, the Wechsler Adult Intelligence Scale Fourth Edition IQ. In addition to requiring only 17–20 of the original 50 words (suggesting that this test could be revised to be up to 66% shorter) and minimizing mean prediction error, the algorithm increases the proportion of the variance in the predicted variable explained in comparison to using all words (from r2 = 0.46 to r2 = 0.61). In a clinical setting this would improve estimates of premorbid cognitive function and, if an abbreviated revision to this test were to be adopted, reduce the arduousness of the test for patients. The proposed method is evaluated with jackknifing and leave one out cross validation. The general approach may be used to optimize the relationship between any two psychological tests by finding the question subset in one test that minimizes the prediction error in a second test by training the genetic algorithm using data collected from participants upon whom both tests have been administered. This approach may also be used to develop new predictive tests, since it provides a method to identify an optimal subset of a set of candidate questions (for which empirical data have been collected) that maximizes prediction accuracy and the proportion of variance in the predicted variable that can be explained. [ABSTRACT FROM AUTHOR]

In palynology, the visual classification of pollen grains from different species is a hard task which is usually tackled by human operators using microscopes. Its complete automatization would save a high quantity of resources and provide valuable improvements especially for allergy-related information systems, but also for other application fields as paleoclimate reconstruction, quality control of honey based products, collection of evidences in criminal investigations or fabric dating and tracking. This paper presents three state-of-the-art deep learning classification methods applied to the recently published POLEN23E image dataset. The three methods make use of convolutional neural networks: the first one is strictly based on the idea of transfer learning, the second one is based on feature extraction and the third one represents a hybrid approach, combining transfer learning and feature extraction. The results from the three methods are indeed very good, reaching over 97% correct classification rates in images not previously seen by the models, where other authors reported around 70. [ABSTRACT FROM AUTHOR]

Feature selection and classification are the main topics in microarray data analysis. Although many feature selection methods have been proposed and developed in this field, SVM-RFE (Support Vector Machine based on Recursive Feature Elimination) is proved as one of the best feature selection methods, which ranks the features (genes) by training support vector machine classification model and selects key genes combining with recursive feature elimination strategy. The principal drawback of SVM-RFE is the huge time consumption. To overcome this limitation, we introduce a more efficient implementation of linear support vector machines and improve the recursive feature elimination strategy and then combine them together to select informative genes. Besides, we propose a simple resampling method to preprocess the datasets, which makes the information distribution of different kinds of samples balanced and the classification results more credible. Moreover, the applicability of four common classifiers is also studied in this paper. Extensive experiments are conducted on six most frequently used microarray datasets in this field, and the results show that the proposed methods have not only reduced the time consumption greatly but also obtained comparable classification performance. [ABSTRACT FROM AUTHOR]

This paper addresses the overlearning problem in the independent component analysis (ICA) used for the removal of muscular artifacts from electroencephalographic (EEG) records. We note that for short EEG records with high number of channels the ICA fails to separate artifact-free EEG and muscular artifacts, which has been previously attributed to the phenomenon called overlearning. We address this problem by projecting an EEG record into several subspaces with a lower dimension, and perform the ICA on each subspace separately. Due to a reduced dimension of the subspaces, the overlearning is suppressed, and muscular artifacts are better separated. Once the muscular artifacts are removed, the signals in the individual subspaces are combined to provide an artifact free EEG record. We show that for short signals and high number of EEG channels our approach outperforms the currently available ICA based algorithms for muscular artifact removal. The proposed technique can efficiently suppress ICA overlearning for short signal segments of high density EEG signals. [ABSTRACT FROM AUTHOR]

This paper addresses a chaos kernel function for the relevance vector machine (RVM) in EEG signal classification, which is an important component of Brain-Computer Interface (BCI). The novel kernel function has evolved from a chaotic system, which is inspired by the fact that human brain signals depict some chaotic characteristics and behaviors. By introducing the chaotic dynamics to the kernel function, the RVM will be enabled for higher classification capacity. The proposed method is validated within the framework of one versus one common spatial pattern (OVO-CSP) classifier to classify motor imagination (MI) of four movements in a public accessible dataset. To illustrate the performance of the proposed kernel function, Gaussian and Polynomial kernel functions are considered for comparison. Experimental results show that the proposed kernel function achieved higher accuracy than Gaussian and Polynomial kernel functions, which shows that the chaotic behavior consideration is helpful in the EEG signal classification. [ABSTRACT FROM AUTHOR]

Aiming to solve the safety problems of high-speed railway operation and management, one new method is urgently needed to construct on the basis of the rough set theory and the uncertainty measurement theory. The method should carefully consider every factor of high-speed railway operation that realizes the measurement indexes of its safety operation. After analyzing the factors that influence high-speed railway safety operation in detail, a rough measurement model is finally constructed to describe the operation process. Based on the above considerations, this paper redistricts the safety influence factors of high-speed railway operation as 16 measurement indexes which include staff index, vehicle index, equipment index and environment. And the paper also provides another reasonable and effective theoretical method to solve the safety problems of multiple attribute measurement in high-speed railway operation. As while as analyzing the operation data of 10 pivotal railway lines in China, this paper respectively uses the rough set-based measurement model and value function model (one model for calculating the safety value) for calculating the operation safety value. The calculation result shows that the curve of safety value with the proposed method has smaller error and greater stability than the value function method’s, which verifies the feasibility and effectiveness. [ABSTRACT FROM AUTHOR]

Pruning is applied in order to combat over-fitting problem where the tree is pruned back with the goal of identifying decision tree with the lowest error rate on previously unobserved instances, breaking ties in favour of smaller trees with high accuracy. In this paper, pruning with Bayes minimum risk is introduced for estimating the risk-rate. This method proceeds in a bottom-up fashion converting a parent node of a subtree to a leaf node if the estimated risk-rate of the parent node for that subtree is less than the risk-rates of its leaf. This paper proposes a post-pruning method that considers various evaluation standards such as attribute selection, accuracy, tree complexity, and time taken to prune the tree, precision/recall scores, TP/FN rates and area under ROC. The experimental results show that the proposed method produces better classification accuracy and its complexity is not much different than the complexities of reduced-error pruning and minimum-error pruning approaches. The experiments also demonstrate that the proposed method shows satisfactory performance in terms of precision score, recall score, TP rate, FP rate and area under ROC. [ABSTRACT FROM AUTHOR]

This paper aims to reschedule online metro trains in delay scenarios. A graph representation and a mixed integer programming model are proposed to formulate the optimization problem. The solution approach is a two-stage optimization method. In the first stage, based on a proposed train state graph and system analysis, the primary and flow-on delays are specifically analyzed and identified with a critical path algorithm. For the second stage a hybrid genetic algorithm is designed to optimize the schedule, with the delay identification results as input. Then, based on the infrastructure data of Beijing Subway Line 4 of China, case studies are presented to demonstrate the effectiveness and efficiency of the solution approach. The results show that the algorithm can quickly and accurately identify primary delays among different types of delays. The economic cost of energy consumption and total delay is considerably reduced (by more than 10% in each case). The computation time of the Hybrid-GA is low enough for rescheduling online. Sensitivity analyses further demonstrate that the proposed approach can be used as a decision-making support tool for operators. [ABSTRACT FROM AUTHOR]

In this paper, we present a new method to recognise the leaf type and identify plant species using phenetic parts of the leaf; lobes, apex and base detection. Most of the research in this area focuses on the popular features such as the shape, colour, vein, and texture, which consumes large amounts of computational processing and are not efficient, especially in the Acer database with a high complexity structure of the leaves. This paper is focused on phenetic parts of the leaf which increases accuracy. Detecting the local maxima and local minima are done based on Centroid Contour Distance for Every Boundary Point, using north and south region to recognise the apex and base. Digital morphology is used to measure the leaf shape and the leaf margin. Centroid Contour Gradient is presented to extract the curvature of leaf apex and base. We analyse 32 leaf images of tropical plants and evaluated with two different datasets, Flavia, and Acer. The best accuracy obtained is 94.76% and 82.6% respectively. Experimental results show the effectiveness of the proposed technique without considering the commonly used features with high computational cost. [ABSTRACT FROM AUTHOR]

Traditional saliency detection algorithms lack object semantic character, and the segmentation algorithms cannot highlight the saliency of the segmentation regions. In order to compensate for the defects of these two algorithms, the salient object segmentation model, which is a novel combination of two algorithms, is established in this paper. With the help of a priori knowledge of image boundary background traits, the K-means++ algorithm is used to cluster the pixels for each region; in line with the sensitivity of the human eye to color and with its attention mechanism, the joint probability distribution of the regional contrast ratio and spatial saliency is established. The selection of the salient area is based on the probabilities, for which the region boundary is taken as the initial curve, and the level-set algorithm is used to perform the salient object segmentation of the image. The curve convergence condition is established according to the confidence level for the segmented region, thus avoiding over-convergence of the segmentation curve. With this method, the salient region boundary is adjacent to the object contour, so the curve evolution time is shorter, and compared with the traditional Li algorithm, the proposed algorithm has higher segmentation evaluation scores, with the additional benefit of emphasizing the importance of the object. [ABSTRACT FROM AUTHOR]

A long-term goal of AI is to produce agents that can learn a diversity of skills throughout their lifetimes and continuously improve those skills via experience. A longstanding obstacle towards that goal is catastrophic forgetting, which is when learning new information erases previously learned information. Catastrophic forgetting occurs in artificial neural networks (ANNs), which have fueled most recent advances in AI. A recent paper proposed that catastrophic forgetting in ANNs can be reduced by promoting modularity, which can limit forgetting by isolating task information to specific clusters of nodes and connections (functional modules). While the prior work did show that modular ANNs suffered less from catastrophic forgetting, it was not able to produce ANNs that possessed task-specific functional modules, thereby leaving the main theory regarding modularity and forgetting untested. We introduce diffusion-based neuromodulation, which simulates the release of diffusing, neuromodulatory chemicals within an ANN that can modulate (i.e. up or down regulate) learning in a spatial region. On the simple diagnostic problem from the prior work, diffusion-based neuromodulation 1) induces task-specific learning in groups of nodes and connections (task-specific localized learning), which 2) produces functional modules for each subtask, and 3) yields higher performance by eliminating catastrophic forgetting. Overall, our results suggest that diffusion-based neuromodulation promotes task-specific localized learning and functional modularity, which can help solve the challenging, but important problem of catastrophic forgetting. [ABSTRACT FROM AUTHOR]

In an age when people are predisposed to report real-world events through their social media accounts, many researchers value the benefits of mining user generated content from social media. Compared with the traditional news media, social media services, such as Twitter, can provide more complete and timely information about the real-world events. However events are often like a puzzle and in order to solve the puzzle/understand the event, we must identify all the sub-events or pieces. Existing Twitter event monitoring systems for sub-event detection and summarization currently typically analyse events based on partial data as conventional data collection methodologies are unable to collect comprehensive event data. This results in existing systems often being unable to report sub-events in real-time and often in completely missing sub-events or pieces in the broader event puzzle. This paper proposes a Sub-event detection by real-TIme Microblog monitoring (STRIM) framework that leverages the temporal feature of an expanded set of news-worthy event content. In order to more comprehensively and accurately identify sub-events this framework first proposes the use of adaptive microblog crawling. Our adaptive microblog crawler is capable of increasing the coverage of events while minimizing the amount of non-relevant content. We then propose a stream division methodology that can be accomplished in real time so that the temporal features of the expanded event streams can be analysed by a burst detection algorithm. In the final steps of the framework, the content features are extracted from each divided stream and recombined to provide a final summarization of the sub-events. The proposed framework is evaluated against traditional event detection using event recall and event precision metrics. Results show that improving the quality and coverage of event contents contribute to better event detection by identifying additional valid sub-events. The novel combination of our proposed adaptive crawler and our stream division/recombination technique provides significant gains in event recall (44.44%) and event precision (9.57%). The addition of these sub-events or pieces, allows us to get closer to solving the event puzzle. [ABSTRACT FROM AUTHOR]

Vagus nerve stimulation (VNS) is an established adjunctive therapy for pharmacologically refractory epilepsy and depression and is currently in active clinical research for other applications. In current clinical studies, VNS is delivered in an open-loop approach, where VNS parameters are defined during a manual titration phase. However, the physiological response to a given VNS configuration shows significant inter and intra-patient variability and may significantly evolve through time. VNS closed-loop approaches, allowing for the optimization of the therapy in an adaptive manner, may be necessary to improve efficacy while reducing side effects. This paper proposes a generic, closed-loop control VNS system that is able to optimize a number of VNS parameters in an adaptive fashion, in order to keep a control variable within a specified range. Although the proposed control method is completely generic, an example application using the cardiac beat to beat interval (RR) as control variable will be developed in this paper. The proposed controller is based on a state transition model (STM) that can be configured using a partially or a fully-connected architecture, different model orders and different state-transition algorithms. The controller is applied to the adaptive regulation of heart rate and evaluated on 6 sheep, for 13 different targets, using partially-connected STM with 10 states. Also, partially and fully-connected STM defined by 30 states were applied to 7 other sheep for the same 10 targets. Results illustrate the interest of the proposed fully-connected STM and the feasibility of integrating this control system into an implantable neuromodulator. [ABSTRACT FROM AUTHOR]

Hand prostheses controlled by surface electromyography are promising due to the non-invasive approach and the control capabilities offered by machine learning. Nevertheless, dexterous prostheses are still scarcely spread due to control difficulties, low robustness and often prohibitive costs. Several sEMG acquisition setups are now available, ranging in terms of costs between a few hundred and several thousand dollars. The objective of this paper is the relative comparison of six acquisition setups on an identical hand movement classification task, in order to help the researchers to choose the proper acquisition setup for their requirements. The acquisition setups are based on four different sEMG electrodes (including Otto Bock, Delsys Trigno, Cometa Wave + Dormo ECG and two Thalmic Myo armbands) and they were used to record more than 50 hand movements from intact subjects with a standardized acquisition protocol. The relative performance of the six sEMG acquisition setups is compared on 41 identical hand movements with a standardized feature extraction and data analysis pipeline aimed at performing hand movement classification. Comparable classification results are obtained with three acquisition setups including the Delsys Trigno, the Cometa Wave and the affordable setup composed of two Myo armbands. The results suggest that practical sEMG tests can be performed even when costs are relevant (e.g. in small laboratories, developing countries or use by children). All the presented datasets can be used for offline tests and their quality can easily be compared as the data sets are publicly available. [ABSTRACT FROM AUTHOR]

Many countries have some kind of energy-system transformation either planned or ongoing for various reasons, such as to curb carbon emissions or to compensate for the phasing out of nuclear energy. One important component of these transformations is the overall reduction in energy demand. It is generally acknowledged that the domestic sector represents a large share of total energy consumption in many countries. Increased energy efficiency is one factor that reduces energy demand, but behavioral approaches (known as “sufficiency”) and their respective interventions also play important roles. In this paper, we address citizens’ heterogeneity regarding both their current behaviors and their willingness to realize their sufficiency potentials—that is, to reduce their energy consumption through behavioral change. We collaborated with three Swiss cities for this study. A survey conducted in the three cities yielded thematic sets of energy-consumption behavior that various groups of participants rated differently. Using this data, we identified four groups of participants with different patterns of both current behaviors and sufficiency potentials. The paper discusses intervention types and addresses citizens’ heterogeneity and behaviors from a city-based perspective. [ABSTRACT FROM AUTHOR]

Urban spaces have a great impact on how people’s emotion and behaviour. There are number of factors that impact our brain responses to a space. This paper presents a novel urban place recommendation approach, that is based on modelling in-situ EEG data. The research investigations leverages on newly affordable Electroencephalogram (EEG) headsets, which has the capability to sense mental states such as meditation and attention levels. These emerging devices have been utilized in understanding how human brains are affected by the surrounding built environments and natural spaces. In this paper, mobile EEG headsets have been used to detect mental states at different types of urban places. By analysing and modelling brain activity data, we were able to classify three different places according to the mental state signature of the users, and create an association map to guide and recommend people to therapeutic places that lessen brain fatigue and increase mental rejuvenation. Our mental states classifier has achieved accuracy of (%90.8). NeuroPlace breaks new ground not only as a mobile ubiquitous brain monitoring system for urban computing, but also as a system that can advise urban planners on the impact of specific urban planning policies and structures. We present and discuss the challenges in making our initial prototype more practical, robust, and reliable as part of our on-going research. In addition, we present some enabling applications using the proposed architecture. [ABSTRACT FROM AUTHOR]

This paper presents the use of the Aimsun microsimulation program to simulate vehicle violating behaviors and observe their impact on road traffic crash risk. Plugins for violations of speeding, slow driving, and abrupt stopping were developed using Aimsun’s API and SDK module. A safety analysis plugin for investigating probability of rear-end collisions was developed, and a method for analyzing collision risk is proposed. A Fuzzy C-mean Clustering algorithm was developed to identify high risk states in different road segments over time. Results of a simulation experiment based on the G15 Expressway in Shanghai showed that abrupt stopping had the greatest impact on increasing collision risk, and the impact of violations increased with traffic volume. The methodology allows for the evaluation and monitoring of risks, alerting of road hazards, and identification of hotspots, and could be applied to the operations of existing facilities or planning of future ones. [ABSTRACT FROM AUTHOR]

This paper presents a stable and fast algorithm for independent component analysis with reference (ICA-R). This is a technique for incorporating available reference signals into the ICA contrast function so as to form an augmented Lagrangian function under the framework of constrained ICA (cICA). The previous ICA-R algorithm was constructed by solving the optimization problem via a Newton-like learning style. Unfortunately, the slow convergence and potential misconvergence limit the capability of ICA-R. This paper first investigates and probes the flaws of the previous algorithm and then introduces a new stable algorithm with a faster convergence speed. There are two other highlights in this paper: first, new approaches, including the reference deflation technique and a direct way of obtaining references, are introduced to facilitate the application of ICA-R; second, a new method is proposed that the new ICA-R is used to recover the complete underlying sources with new advantages compared with other classical ICA methods. Finally, the experiments on both synthetic and real-world data verify the better performance of the new algorithm over both previous ICA-R and other well-known methods. [ABSTRACT FROM AUTHOR]

In big data area a significant challenge about string similarity join is to find all similar pairs more efficiently. In this paper, we propose a parallel processing framework for efficient string similarity join. First, the input is split into some disjoint small subsets according to the joint frequency distribution and the interval distribution of strings. Then the filter-verification strategy is adopted in the computation of string similarity for each subset so that the number of candidate pairs is reduced before an effective pruning strategy is used to improve the performance. Finally, the operation of string join is executed in parallel. Para-Join algorithm based on the multi-threading technique is proposed to implement the framework in a multi-core system while Pada-Join algorithm based on Spark platform is proposed to implement the framework in a cluster system. We prove that Para-Join and Pada-Join cannot only avoid reduplicate computation but also ensure the completeness of the result. Experimental results show that Para-Join can achieve high efficiency and significantly outperform than state-of-the-art approaches, meanwhile, Pada-Join can work on large datasets. [ABSTRACT FROM AUTHOR]

Among Parkinson’s disease (PD) symptoms, freezing of gait (FoG) is one of the most debilitating. To assess FoG, current clinical practice mostly employs repeated evaluations over weeks and months based on questionnaires, which may not accurately map the severity of this symptom. The use of a non-invasive system to monitor the activities of daily living (ADL) and the PD symptoms experienced by patients throughout the day could provide a more accurate and objective evaluation of FoG in order to better understand the evolution of the disease and allow for a more informed decision-making process in making adjustments to the patient’s treatment plan. This paper presents a new algorithm to detect FoG with a machine learning approach based on Support Vector Machines (SVM) and a single tri-axial accelerometer worn at the waist. The method is evaluated through the acceleration signals in an outpatient setting gathered from 21 PD patients at their home and evaluated under two different conditions: first, a generic model is tested by using a leave-one-out approach and, second, a personalised model that also uses part of the dataset from each patient. Results show a significant improvement in the accuracy of the personalised model compared to the generic model, showing enhancement in the specificity and sensitivity geometric mean (GM) of 7.2%. Furthermore, the SVM approach adopted has been compared to the most comprehensive FoG detection method currently in use (referred to as MBFA in this paper). Results of our novel generic method provide an enhancement of 11.2% in the GM compared to the MBFA generic model and, in the case of the personalised model, a 10% of improvement with respect to the MBFA personalised model. Thus, our results show that a machine learning approach can be used to monitor FoG during the daily life of PD patients and, furthermore, personalised models for FoG detection can be used to improve monitoring accuracy. [ABSTRACT FROM AUTHOR]

Breast cancer is an all too common disease in women, making how to effectively predict it an active research problem. A number of statistical and machine learning techniques have been employed to develop various breast cancer prediction models. Among them, support vector machines (SVM) have been shown to outperform many related techniques. To construct the SVM classifier, it is first necessary to decide the kernel function, and different kernel functions can result in different prediction performance. However, there have been very few studies focused on examining the prediction performances of SVM based on different kernel functions. Moreover, it is unknown whether SVM classifier ensembles which have been proposed to improve the performance of single classifiers can outperform single SVM classifiers in terms of breast cancer prediction. Therefore, the aim of this paper is to fully assess the prediction performance of SVM and SVM ensembles over small and large scale breast cancer datasets. The classification accuracy, ROC, F-measure, and computational times of training SVM and SVM ensembles are compared. The experimental results show that linear kernel based SVM ensembles based on the bagging method and RBF kernel based SVM ensembles with the boosting method can be the better choices for a small scale dataset, where feature selection should be performed in the data pre-processing stage. For a large scale dataset, RBF kernel based SVM ensembles based on boosting perform better than the other classifiers. [ABSTRACT FROM AUTHOR]

Customized buses (CBs) are a new mode of public transportation and an important part of diversified public transportation, providing advanced, attractive and user-led service. The operational activity of a CB is planned by aggregating space–time demand and similar passenger travel demands. Based on an analysis of domestic and international research and the current development of CBs in China and considering passenger travel data, this paper studies the problems associated with the operation of CBs, such as stop selection, line planning and timetables, and establishes a model for the stop planning and timetables of CBs. The improved immune genetic algorithm (IIGA) is used to solve the model with regard to the following: 1) multiple population design and transport operator design, 2) memory library design, 3) mutation probability design and crossover probability design, and 4) the fitness calculation of the gene segment. Finally, a real-world example in Beijing is calculated, and the model and solution results are verified and analyzed. The results illustrate that the IIGA solves the model and is superior to the basic genetic algorithm in terms of the number of passengers, travel time, average passenger travel time, average passenger arrival time ahead of schedule and total line revenue. This study covers the key issues involving operational systems of CBs, combines theoretical research and empirical analysis, and provides a theoretical foundation for the planning and operation of CBs. [ABSTRACT FROM AUTHOR]

For ensemble learning, how to select and combine the candidate classifiers are two key issues which influence the performance of the ensemble system dramatically. Random vector functional link networks (RVFL) without direct input-to-output links is one of suitable base-classifiers for ensemble systems because of its fast learning speed, simple structure and good generalization performance. In this paper, to obtain a more compact ensemble system with improved convergence performance, an improved ensemble of RVFL based on attractive and repulsive particle swarm optimization (ARPSO) with double optimization strategy is proposed. In the proposed method, ARPSO is applied to select and combine the candidate RVFL. As for using ARPSO to select the optimal base RVFL, ARPSO considers both the convergence accuracy on the validation data and the diversity of the candidate ensemble system to build the RVFL ensembles. In the process of combining RVFL, the ensemble weights corresponding to the base RVFL are initialized by the minimum norm least-square method and then further optimized by ARPSO. Finally, a few redundant RVFL is pruned, and thus the more compact ensemble of RVFL is obtained. Moreover, in this paper, theoretical analysis and justification on how to prune the base classifiers on classification problem is presented, and a simple and practically feasible strategy for pruning redundant base classifiers on both classification and regression problems is proposed. Since the double optimization is performed on the basis of the single optimization, the ensemble of RVFL built by the proposed method outperforms that built by some single optimization methods. Experiment results on function approximation and classification problems verify that the proposed method could improve its convergence accuracy as well as reduce the complexity of the ensemble system. [ABSTRACT FROM AUTHOR]

Behaviour models are the most commonly used input for predicting the reliability of a software system at the early design stage. A component behaviour model reveals the structure and behaviour of the component during the execution of system-level functionalities. There are various challenges related to component reliability prediction at the early design stage based on behaviour models. For example, most of the current reliability techniques do not provide fine-grained sequential behaviour models of individual components and fail to consider the loop entry and exit points in the reliability computation. Moreover, some of the current techniques do not tackle the problem of operational data unavailability and the lack of analysis results that can be valuable for software architects at the early design stage. This paper proposes a reliability prediction technique that, pragmatically, synthesizes system behaviour in the form of a state machine, given a set of scenarios and corresponding constraints as input. The state machine is utilized as a base for generating the component-relevant operational data. The state machine is also used as a source for identifying the nodes and edges of a component probabilistic dependency graph (CPDG). Based on the CPDG, a stack-based algorithm is used to compute the reliability. The proposed technique is evaluated by a comparison with existing techniques and the application of sensitivity analysis to a robotic wheelchair system as a case study. The results indicate that the proposed technique is more relevant at the early design stage compared to existing works, and can provide a more realistic and meaningful prediction. [ABSTRACT FROM AUTHOR]

Pertussis is a contagious respiratory disease which mainly affects young children and can be fatal if left untreated. The World Health Organization estimates 16 million pertussis cases annually worldwide resulting in over 200,000 deaths. It is prevalent mainly in developing countries where it is difficult to diagnose due to the lack of healthcare facilities and medical professionals. Hence, a low-cost, quick and easily accessible solution is needed to provide pertussis diagnosis in such areas to contain an outbreak. In this paper we present an algorithm for automated diagnosis of pertussis using audio signals by analyzing cough and whoop sounds. The algorithm consists of three main blocks to perform automatic cough detection, cough classification and whooping sound detection. Each of these extract relevant features from the audio signal and subsequently classify them using a logistic regression model. The output from these blocks is collated to provide a pertussis likelihood diagnosis. The performance of the proposed algorithm is evaluated using audio recordings from 38 patients. The algorithm is able to diagnose all pertussis successfully from all audio recordings without any false diagnosis. It can also automatically detect individual cough sounds with 92% accuracy and PPV of 97%. The low complexity of the proposed algorithm coupled with its high accuracy demonstrates that it can be readily deployed using smartphones and can be extremely useful for quick identification or early screening of pertussis and for infection outbreaks control. [ABSTRACT FROM AUTHOR]

Compressive sensing (CS) theory asserts that we can reconstruct signals and images with only a small number of samples or measurements. Recent works exploiting the nonlocal similarity have led to better results in various CS studies. To better exploit the nonlocal similarity, in this paper, we propose a non-convex smoothed rank function based model for CS image reconstruction. We also propose an efficient alternating minimization method to solve the proposed model, which reduces a difficult and coupled problem to two tractable subproblems. Experimental results have shown that the proposed method performs better than several existing state-of-the-art CS methods for image reconstruction. [ABSTRACT FROM AUTHOR]

Information confidentiality is an essential requirement for cyber security in critical infrastructure. Identity-based cryptography, an increasingly popular branch of cryptography, is widely used to protect the information confidentiality in the critical infrastructure sector due to the ability to directly compute the user’s public key based on the user’s identity. However, computational requirements complicate the practical application of Identity-based cryptography. In order to improve the efficiency of identity-based cryptography, this paper presents an effective method to construct pairing-friendly elliptic curves with low hamming weight 4 under embedding degree 1. Based on the analysis of the Complex Multiplication(CM) method, the soundness of our method to calculate the characteristic of the finite field is proved. And then, three relative algorithms to construct pairing-friendly elliptic curve are put forward. 10 elliptic curves with low hamming weight 4 under 160 bits are presented to demonstrate the utility of our approach. Finally, the evaluation also indicates that it is more efficient to compute Tate pairing with our curves, than that of Bertoni et al. [ABSTRACT FROM AUTHOR]

In this paper we apply machine learning methods for predicting protein interactions in fungal secretion pathways. We assume an inter-species transfer setting, where training data is obtained from a single species and the objective is to predict protein interactions in other, related species. In our methodology, we combine several state of the art machine learning approaches, namely, multiple kernel learning (MKL), pairwise kernels and kernelized structured output prediction in the supervised graph inference framework. For MKL, we apply recently proposed centered kernel alignment and p-norm path following approaches to integrate several feature sets describing the proteins, demonstrating improved performance. For graph inference, we apply input-output kernel regression (IOKR) in supervised and semi-supervised modes as well as output kernel trees (OK3). In our experiments simulating increasing genetic distance, Input-Output Kernel Regression proved to be the most robust prediction approach. We also show that the MKL approaches improve the predictions compared to uniform combination of the kernels. We evaluate the methods on the task of predicting protein-protein-interactions in the secretion pathways in fungi, S.cerevisiae, baker’s yeast, being the source, T. reesei being the target of the inter-species transfer learning. We identify completely novel candidate secretion proteins conserved in filamentous fungi. These proteins could contribute to their unique secretion capabilities. [ABSTRACT FROM AUTHOR]

Recent research has demonstrated that characteristic gene selection based on gene expression data remains faced with considerable challenges. This is primarily because gene expression data are typically high dimensional, negative, non-sparse and noisy. However, existing methods for data analysis are able to cope with only some of these challenges. In this paper, we address all of these challenges with a unified method: nonnegative matrix factorization via the L2,1-norm (NMF-L2,1). While L2,1-norm minimization is applied to both the error function and the regularization term, our method is robust to outliers and noise in the data and generates sparse results. The application of our method to plant and tumor gene expression data demonstrates that NMF-L2,1 can extract more characteristic genes than other existing state-of-the-art methods. [ABSTRACT FROM AUTHOR]

As a common ocular complication for diabetic patients, diabetic retinopathy has become an important public health problem in the world. Early diagnosis and early treatment with the help of fundus imaging technology is an effective control method. In this paper, a robust inverse diffusion equation combining a self-similarity filtering is presented to detect and evaluate diabetic retinopathy using retinal image enhancement. A flux corrected transport technique is used to control diffusion flux adaptively, which eliminates overshoots inherent in the Laplacian operation. Feature preserving denoising by the self-similarity filtering ensures a robust enhancement of noisy and blurry retinal images. Experimental results demonstrate that this algorithm can enhance important details of retinal image data effectively, affording an opportunity for better medical interpretation and subsequent processing. [ABSTRACT FROM AUTHOR]

The identification of a subset of genes having the ability to capture the necessary information to distinguish classes of patients is crucial in bioinformatics applications. Ensemble and bagging methods have been shown to work effectively in the process of gene selection and classification. Testament to that is random forest which combines random decision trees with bagging to improve overall feature selection and classification accuracy. Surprisingly, the adoption of these methods in support vector machines has only recently received attention but mostly on classification not gene selection. This paper introduces an ensemble SVM-Recursive Feature Elimination (ESVM-RFE) for gene selection that follows the concepts of ensemble and bagging used in random forest but adopts the backward elimination strategy which is the rationale of RFE algorithm. The rationale behind this is, building ensemble SVM models using randomly drawn bootstrap samples from the training set, will produce different feature rankings which will be subsequently aggregated as one feature ranking. As a result, the decision for elimination of features is based upon the ranking of multiple SVM models instead of choosing one particular model. Moreover, this approach will address the problem of imbalanced datasets by constructing a nearly balanced bootstrap sample. Our experiments show that ESVM-RFE for gene selection substantially increased the classification performance on five microarray datasets compared to state-of-the-art methods. Experiments on the childhood leukaemia dataset show that an average 9% better accuracy is achieved by ESVM-RFE over SVM-RFE, and 5% over random forest based approach. The selected genes by the ESVM-RFE algorithm were further explored with Singular Value Decomposition (SVD) which reveals significant clusters with the selected data. [ABSTRACT FROM AUTHOR]