Search

Current emotion recognition computational techniques have been successful on associating the emotional changes with the EEG signals, and so they can be identified and classified from EEG signals if appropriate stimuli are applied. However, automatic recognition is usually restricted to a small number of emotions classes mainly due to signal’s features and noise, EEG constraints and subject-dependent issues. In order to address these issues, in this paper a novel feature-based emotion recognition model is proposed for EEG-based Brain–Computer Interfaces. Unlike other approaches, our method explores a wider set of emotion types and incorporates additional features which are relevant for signal pre-processing and recognition classification tasks, based on a dimensional model of emotions: Valence and Arousal . It aims to improve the accuracy of the emotion classification task by combining mutual information based feature selection methods and kernel classifiers. Experiments using our approach for emotion classification which combines efficient feature selection methods and efficient kernel-based classifiers on standard EEG datasets show the promise of the approach when compared with state-of-the-art computational methods. [ABSTRACT FROM AUTHOR]

In this paper, I investigate the phenomenon of ‘verb classifiers’, i.e. nominal classifiers that attach to the verb and that classify one of its arguments. A sample of thirteen supposed ‘verb classifier’ languages is presented and analysed in detail; the elicited data argues against the maintenance of the general theoretical notion of an independent classifier type of ‘verb classifiers’, since the analysed systems either modify the concept expressed by the predicate (and thus do not classify a verbal argument) or mark a verbal argument instead of classifying it. [ABSTRACT FROM AUTHOR]

Binary classification is a common statistical learning problem in which a model is estimated on a set of covariates for some outcome, indicating the membership of one of two classes. In the literature, there exists a distinction between hard and soft classification. In soft classification, the conditional class probability is modeled as a function of the covariates. In contrast, hard classification methods only target the optimal prediction boundary. While hard and soft classification methods have been studied extensively, not much work has been performed to compare the actual tasks of hard and soft classification. In this paper, we propose a spectrum of statistical learning problems that span the hard and soft classification tasks based on fitting multiple decision rules to the data. By doing so, we reveal a novel collection of learning tasks of increasing complexity. We study the problems using the framework of large-margin classifiers and a class of piecewise linear convex surrogates, for which we derive statistical properties and a corresponding sub-gradient descent algorithm. We conclude by applying our approach to simulation settings and a magnetic resonance imaging (MRI) dataset from the Alzheimer's Disease Neuroimaging Initiative (ADNI) study. [ABSTRACT FROM AUTHOR]

This paper attempts to compare and analyze numeral classifiers of Bodo, Rabha and Kok Borok languages. They have been recognised with same sub-group and they migrated together in North-East India. (Chatterji, Suniti Kumar, Kirata Jana Kriti, 2007, page -45-46). Some Similarities and Dissimilarities are found among these languages. Each language has its own Structures and Functions in numeral classifiers. [ABSTRACT FROM AUTHOR]

Automatic text classification is a prominent research topic in text mining. The text pre-processing is a major role in text classifier. The efficiency of pre-processing techniques is increasing the performance of text classifier. In this paper, we are implementing ECAS stemmer, Efficient Instance Selection and Pre-computed Kernel Support Vector Machine for text classification using recent research articles. We are using better pre-processing techniques such as ECAS stemmer to find root word, Efficient Instance Selection for dimensionality reduction of text data and Pre-computed Kernel Support Vector Machine for classification of selected instances. In this experiments were performed on 750 research articles with three classes such as engineering article, medical articles and educational articles. The EIS-SVM classifier provides better performance in real-time research articles classification. [ABSTRACT FROM AUTHOR]

k NN is extensively studied for multilabel classification in the literature. Several k NN-based multilabel learning algorithms have been witnessed during the past years. They usually take k NN as their base classifiers to construct classification models, and then predict the class labels by virtue of Bayesian or majority rules. In this paper, a nearest neighbor selection for multilabel classification is proposed. Specifically, the target labels of new data are predicted with the help of those relevant and reliable data, which explored by the concept of shelly nearest neighbor. For effectiveness, the certainty factor is further adopted to well address the problem of unbalanced and uncertain data. The comparison experiments with eleven popular multilabel classifiers are conducted on ten benchmark data sets. The experimental results show that the performance of the proposed method is competitive and outperforms the popular multilabel classifiers in most cases. [ABSTRACT FROM AUTHOR]

Spectral imputation and classifier modification can be counted as the two main missing data approaches for robust automatic speech recognition (ASR). Despite their potentials, little attention has been paid to the classifier modification techniques. In this paper, we show that transferring bounded marginalization, which is a classifier modification method, from spectral to cepstral domain would be beneficial for robust ASR. We also propose improved solutions on this transfer toward a better performance. Two such techniques are presented. The first approach still does not need training of any extra model. It benefits from an observed characteristic of cepstral features and raises accuracy of previously proposed method to a comparable level with that of a classic imputation method. The second technique combines our originally proposed method with an imputation technique but replaces spectral reconstruction with a simpler and faster possible range estimation of missing components. We show that the resulting method improves the accuracies of either of the two combined methods. The proposed techniques also show good robustness when implemented with an inaccurate spectrographic mask. [ABSTRACT FROM AUTHOR]

We propose a method for pedestrian detection from aerial images captured by unmanned aerial vehicles (UAVs). Aerial images are captured at considerably low resolution, and they are often subject to heavy noise and blur as a result of atmospheric influences. Furthermore, significant changes to the appearance of pedestrians frequently occur because of UAV motion. In order to address these crucial problems, we propose a cascading classifier that concatenates a pre-trained classifier and an online learning-based classifier. We construct the first classifier using deep belief network (DBN) with an extended input layer. Unlike previous approaches that use raw images as the input layer of the DBN, we exploit multi-scale histogram of oriented gradients (MS-HOG) features. The MS-HOG enables us to supply better and richer information than low-resolution aerial images for constructing a reliable deep structure of DBN, because the dimensions of the input features can be expanded. Furthermore, the MS-HOG effectively extracts the necessary edge information while reducing trivial gradients and noise. The second classifier is based on online learning, and it uses predictions of the target appearance using UAV motions. Predicting the target appearance enables us to collect reliable training samples for the classifier's online learning process. Experiments using aerial videos demonstrate the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Adaptive discriminative tracking is a new research topic that has attracted broad attention due to its extensive application value. To take full advantage of the information about targets and their surrounding background, we propose a novel single object tracking-by-detection tracker in this paper, combining semi-supervised learning, multiple instance learning and the Bayesian theorem. The tracker uses a block-based inconsistency function of the labeled and unlabeled training samples in the selection of optimal weak classifiers during the parameter updating phase of each frame. Experimental results showed that the proposed tracker has excellent performance over other eight state-of-the-art trackers for thirteen open-access video sequences. [ABSTRACT FROM AUTHOR]

This paper examines the performance of automatically classifying five tone choices (i.e., falling, rising, rising-falling, falling-rising, and neutral) of Brazil's intonation model. We tested two machine learning classifiers (neural network and boosting ensemble) in two configurations (multi-class and pairwise coupling) and a rule-based classifier. Three sets of acoustic features built from the TILT and Bézier pitch contour models and a new four-point pitch contour model we introduced here were investigated. Tone choices are one of the key elements of Brazil's prosodic intonation model. We found the rule-based classifier, which was built on our four-point model, achieved better results than the others with an accuracy of 75.1 % and a Cohen's kappa coefficient of 0.73. This research proves that it is possible to classify tone choices with an accuracy reaching close to the percentage of agreement between two human analysts. The findings further concluded that our four-point model was better for classifying Brazil's tone choices than both of the TILT or Bézier models. [ABSTRACT FROM AUTHOR]

C.T.C. Wall classified the unimodular complete intersection singularities. He indicated in the list only the -constant strata and not the complete classication in each case. In this article we give a complete list of space curve unimodular singularities and also the description of a classifier. Instead of computing the normal forms, the singularity is identified by certain invariants. [ABSTRACT FROM AUTHOR]

The demographic classification, and gender recognition specifically, is a topic of considerable interest to researchers because of its importance in various applications, such as in areas of surveillance, face recognition, indexing videos, dynamic marketing studies, among other. This paper proposes a new way of carry through gender classification using a frequency vector based on local descriptors SIFT or SURF. The goal is to determine if the frequency vectors contain discriminant information. The training and validation of classification models will be based on Multi-PIE data, which contains face images taken under laboratory conditions, available with illumination changes, pose and expressions. For experimental development only considered images captured with normal room lighting, subjects with neutral expression and 11 changes of pose. The results validated that the proposed models contain discriminant information and also maintain a stable accuracy in gender classification on images with pose variations. The latter is extremely important, because in real life conditions are unlikely to acquire images of frontal faces, rather most will change perspective, rotation and illumination changes, therefore a robust model to these conditions is required. [ABSTRACT FROM AUTHOR]

In recent years, different researchers in the machine learning community have presented new classification frameworks which go beyond the standard supervised classification in different aspects. Specifically, a wide spectrum of novel frameworks that use partially labeled data in the construction of classifiers has been studied. With the objective of drawing up a description of the state-of-the-art, three identifying characteristics of these novel frameworks have been considered: (1) the relationship between instances and labels of a problem, which may be beyond the one-instance one-label standard, (2) the possible provision of partial class information for the training examples, and (3) the possible provision of partial class information also for the examples in the prediction stage. These three ideas have been formulated as axes of a comprehensive taxonomy that organizes the state-of-the-art. The proposed organization allows us both to understand similarities/differences among the different classification problems already presented in the literature as well as to discover unexplored frameworks that might be seen as further challenges and research opportunities. A representative set of state-of-the-art problems has been used to illustrate the novel taxonomy and support the discussion. [ABSTRACT FROM AUTHOR]

This article is an inquiry into how the relationship between the principle of non-contradiction and the limits of thought has been understood by thinkers as diverse as Hegel, Heidegger, Levinas, and Graham Priest. While Heidegger and Levinas focus on the question of temporality and Priest lakes a formal approach, all these philosophers effectively maintain that the principle of non-contradiction imposes a restriction on thought that disables it from adequately accounting for its own limits and thus what lies beyond those limits, the implication being that the violation of the principle is necessary for such an accounting to lake place. However, the ultimate argument here is that, contrary to Priest's interpretation, Hegel's philosophy can be convincingly read as supporting the idea that the mind's ability to go beyond any particular limit of thought can actually be said to involve an adherence to a normative demand to locate and dispel the contradictions that emerge through the very setting of determinative limits. This is a non-formal consistency that evinces a "logic" that is unknowingly followed by the Heideggerian and Levinasian phenomenological philosophies of transcendence. [ABSTRACT FROM AUTHOR]

Wrapper variable selection methods are widely adopted in many applications, among which the design of classifiers. The main problem related to these approaches regards the stability of the selection, namely the exploitation of different training data set can lead to the selection of different variable subsets. This problem is particularly critical in applications where variable selection is used to interpret the behaviour of the process or phenomenon under consideration, i.e. to understand which among a potentially huge list of variables actually affect the classification. The paper proposes a method that improves the stability of the wrapper variable selection procedures while preserving and possibly improving the classification performance. Moreover three binary classifiers are performed in order to prove the effectiveness of the proposed method. [ABSTRACT FROM AUTHOR]

Deep neural networks are currently among the most commonly used classifiers. Despite easily achieving very good performance, one of the best selling points of these models is their modular design – one can conveniently adapt their architecture to specific needs, change connectivity patterns, attach specialised layers, experiment with a large amount of activation functions, normalisation schemes and many others. While one can find impressively wide spread of various configurations of almost every aspect of the deep nets, one element is, in authors’ opinion, underrepresented – while solving classification problems, vast majority of papers and applications simply use log loss. In this paper we try to investigate how particular choices of loss functions affect deep models and their learning dynamics, as well as resulting classifiers robustness to various effects. We perform experiments on classical datasets, as well as provide some additional, theoretical insights into the problem. In particular we show that L1 and L2 losses are, quite surprisingly, justified classification objectives for deep nets, by providing probabilistic interpretation in terms of expected misclassification. We also introduce two losses which are not typically used as deep nets objectives and show that they are viable alternatives to the existing ones. [ABSTRACT FROM AUTHOR]

What do Slavic aspectual prefixes have in common with numeral classifiers? Our answer is that the parallels are compelling, both in terms of breadth and depth. The grammatical function of numeral classifiers is to form and classify units for the referents of nouns, and we argue that Slavic aspectual prefixes have the function of forming and classifying units for the referents of verbs. Numeral classifiers contribute a meaning of discreteness to objects, whereas Slavic aspectual prefixes do the same for events. Just as there are various types of numeral classifiers, there are also various types of Slavic aspectual prefixes. We find that the patterns identified for numeral classifiers are consistently matched by the grammatical behavior of the various types of aspectual prefixes throughout the Slavic linguistic territory. We furthermore anchor this comparison in a variety of ways, taking into account distributional and semantic evidence, and the effects of construal, foregrounding, definiteness, and transnumerality. In the places where this comparison breaks down, the causes are inherent differences between the domain of nouns and the domain of verbs. We suggest that Slavic aspectual prefixes and numeral classifiers should be considered to be verbal and nominal instantiations of a general category of lexico-grammatical unitizers. [ABSTRACT FROM AUTHOR]

We deal with the problem of the multiaspect text categorization which calls for the classification of the documents with respect to two, in a sense, orthogonal sets of categories. We briefly define the problem, mainly referring to our previous work, and study the application of the knearest neighbours algorithm. We propose a new technique meant to enhance the effectiveness of this algorithm when applied to the problem in question. We show some experimental results confirming usefulness of the proposed approach. [ABSTRACT FROM AUTHOR]

An example-based face hallucination system is proposed, in which given a low-resolution facial image, a corresponding high-resolution image is automatically obtained. In practice, such a problem is extremely challenging since it is often the case that two discriminative high-resolution images may have similar low-resolution inputs. To address this issue, this study proposes an ensemble of image feature representations, including various local patch- or block-based representations, a one-dimensional vector image representation, a two-dimensional matrix image representation, and a global matrix image representation. Notably, some of these representations are designed to preserve the global facial geometry of the low-resolution input, while others are designed to preserve the local detailed texture. For each feature representation, a regression function is constructed to synthesize a high-resolution image from the low-resolution input image. The synthesis process is conducted in a layer-by-layer fashion, with the output from one layer serving as the input to the following layer. Importantly, each regression function is associated with a classifier in order to determine which regression functions are required in the synthesis procedure in accordance with the particular characteristics of the input image. Furthermore, these classifiers can also help to deal with the individual ambiguity of system low-resolution inputs. The experimental results show that the proposed framework is capable of synthesizing high-resolution images from low-resolution input images with a wide variety of facial poses, geometry misalignments and facial expressions even when such images are not included within the original training dataset. [ABSTRACT FROM AUTHOR]

Translated language is different from non-translated language. It combines the characteristics of both source and target languages, and so presents a set of hybrid features at different linguistic levels. The present research investigates hybridity features of classifier constructions in translated Chinese with the help of comparable and parallel corpora. According to McEnery and Xiao (2007), classifiers occur twentynine times more frequently in Chinese than in English, and Chinese is regarded as a classifier language, while English is a non-classifier language. The frequencies of single classifiers, such as Temporal classifier (qt) and Verbal classifier (qv), are higher in non-translated Chinese than those in translated Chinese. However, classifier constructions such as "YI GE + Noun" are used more frequently in translated Chinese. "YI GE" can be used as an indefinite article. The present research finds that most of the collocations of "YI GE" in non-translated and translated Chinese are similar, though the frequencies are different. Translated Chinese follows the norms of native Chinese, and this is the result of the normalization tendency in translation. However, translated Chinese still shows hybrid characteristics that are distinct from non-translated Chinese in the collocations of "YI GE". The parallel corpus data shows that "a/an + (adj.) + noun" are the most common equivalents of "YI GE", followed by "another", "one + noun", "the noun", "one", etc. The concordancing results also tell us that "YI GE" is used in translated Chinese when there are no articles used in the English source text. The translators add "YI GE" according to the norms of the target Chinese language, and this addition results in the over-representation of "YI GE" in translated Chinese. "YI GE" constructions in translated Chinese can be divided into four classes: "YI + GE + countable noun", "YI + GE + abstract noun", "YI + GE + verb/adjective" and "YI + GE + complex phrases". The combination of source language interference and target language normalization leads to the hybridity of classifier constructions in translated Chinese. [ABSTRACT FROM AUTHOR]

Feature selection plays an important role in emotion recognition from speech signals because it improves the classification accuracy by choosing the best uncorrelated features. In wrapper method of feature selection, the features are evaluated by a classifier. Features of large dimension will increase the computational complexity of the classifier, and further it will affect the training of classifiers which needs inverse of covariance matrix. We propose a four-stage feature selection method which avoids the problem of curse of dimensionality by the principle of divide and conquer. In the proposed method, the dimension of the feature vector is shortened at any stage in a way that the classifiers, whose training is affected by the large feature dimension, can also be used to evaluate the features. Experimental results show that the four-stage feature selection method improves classification accuracy. Another method to improve classification accuracy is evolved by bringing together several classifiers with a fusion technique. Class-specific multiple classifiers scheme is one such method that improves classification accuracy by combining optimum performance feature set and classifier for each emotional class. In this work, we improve the performance of the class-specific multiple classifiers scheme by embedding the proposed feature selection method in its structure. [ABSTRACT FROM AUTHOR]

In this paper, we examine the performance of automatically detecting Brazil's prominent syllables using five machine learning classifiers and seven sets of features consisting of three features: pitch, intensity, and duration, taken one at time, two at a time, and all three. Prominent syllables are the foundation of Brazil's prosodic intonation model. We found that using pitch, intensity, and duration as features produces the best optimal results. Our findings also revealed that in terms of accuracy, F-measure, and Cohen's kappa coefficient that bagging an ensemble of decision tree learners performed the best (accuracy = 95.9 ± 0.2 %; F-measure = 93.7 ± 0.4; κ = 0.907 ± 0.005). The performance of our current model proves to be significantly better than any other automatic detection software that exists or that of human transcription experts of prosody. [ABSTRACT FROM AUTHOR]

The present paper investigates the relationship between form (morpho-syntax) and function (semantic interpretation) as involved in the expression of countability in the Chinese whpronoun system. We focus on three wh-pronouns: duo-shao 'much-little', duo-shao-ge 'much-little-Classifier', and ji-ge 'how many-Classifier'. Based on the examination of the interpretation and distribution of these wh-phrases in three distinct linguistic contexts (questions, bare conditionals, and negative statements), we generalize that while duo-shao functions as an indeterminate determiner, duo-shao-ge and ji-ge function as two count determiners. This generalization invites the conclusion that the morpho-syntax of these whphrases, particularly, the presence/absence of a classifier, determines and shapes their semantic interpretation in countability. From a more general perspective, the present study suggests that a linguistic form determines its own linguistic functions. [ABSTRACT FROM AUTHOR]

This paper outlines our proposal for a novel metacognitive-based scaffolding classifier, namely recurrent classifier (rClass). rClass is capable of emulating three fundamental pillars of human learning in terms of what-to-learn, how-to-learn, and when-to-learn. The cognitive constituent of rClass is underpinned by a recurrent network based on a generalized version of the Takagi–Sugeno–Kang fuzzy system possessing a local feedback of the rule layer. The main basis of the what-to-learn component relies on the new active learning-based conflict measure. Meanwhile, the when-to-learn learning scenario makes use of the standard sample reserved strategy. The how-to-learn module actualizes the Schema and Scaffolding concepts of cognitive psychology. All learning principles are committed in the single-pass local learning modes and create a plug-and-play learning foundation minimizing additional pre- or post-training phases. The efficacy of rClass has been scrutinized by means of rigorous empirical studies, statistical tests, and benchmarks with state-of-the-art classifiers, which demonstrate the rClass potency in producing reliable classification rates, while retaining low complexity in terms of the rule base burden, computational load, and annotation effort. [ABSTRACT FROM AUTHOR]

Sometimes form-meaning mappings in language are not arbitrary, but iconic: they depict what they represent. Incorporating iconic elements of language into a compositional semantics faces a number of challenges in formal frameworks as evidenced by the lengthy literature in linguistics and philosophy on quotation/direct speech, which iconically portrays the words of another in the form that they were used. This paper compares the well-studied type of iconicity found with verbs of quotation with another form of iconicity common in sign languages: classifier predicates. I argue that these two types of verbal iconicity can, and should, incorporate their iconic elements in the same way using event modification via the notion of a context dependent demonstration. This unified formal account of quotation and classifier predicates predicts that a language might use the same strategy for conveying both, and I argue that this is the case with role shift in American Sign Language. Role shift is used to report others' language and thoughts as well as their actions, and recently has been argued to provide evidence in favor of Kaplanian 'monstrous' indexical expressions. By reimagining role shift as involving either (i) quotation for language demonstrations or (ii) 'body classifier' predicates for action demonstrations, the proposed account eliminates one major argument for these monsters coming from sign languages. Throughout this paper, sign languages provide a fruitful perspective for studying quotation and other forms of iconicity in natural language due to their (i) lack of a commonly used writing system which is otherwise often mistaken as primary data instead of speech, (ii) the rich existing literature on iconicity within sign language linguistics, and (iii) the ability of role shift to overtly mark the scope of a language report. In this view, written language is merely a special case of a more general phenomenon of sign and speech demonstration, which accounts more accurately for natural language data by permitting more strict or loose verbatim interpretations of demonstrations through the context dependent pragmatics. [ABSTRACT FROM AUTHOR]

The present paper comments on signs of American Sign Language in the perspective of relevance theory. The main claim is that classifiers encode procedural instructions to help the addressee pick out the intended referent for the procedural referring expressions made with classifier constructions. The author explains how three classes of classifiers differently manipulate concepts to instruct the addressee to create ad hoc concepts though the use of inference, narrowing, and broadening. It is also claimed that classifier constructions do not encode a conceptual meaning, but a procedural instruction. The discussion includes illustrations of how the speaker's using classifier constructions instead of lexical signs may increase the number of cognitive effects on the part of the addressee. [ABSTRACT FROM AUTHOR]

This paper investigates a relationship between the fuzziness of a classifier and the misclassification rate of the classifier on a group of samples. For a given trained classifier that outputs a membership vector, we demonstrate experimentally that samples with higher fuzziness outputted by the classifier mean a bigger risk of misclassification. We then propose a fuzziness category based divide-and-conquer strategy which separates the high-fuzziness samples from the low fuzziness samples. A particular technique is used to handle the high-fuzziness samples for promoting the classifier performance. The reasonability of the approach is theoretically explained and its effectiveness is experimentally demonstrated. [ABSTRACT FROM AUTHOR]

Discourse parsing has become an inevitable task to process information in the natural language processing arena. Parsing complex discourse structures beyond the sentence level is a significant challenge. This article proposes a discourse parser that constructs rhetorical structure (RS) trees to identify such complex discourse structures. Unlike previous parsers that construct RS trees using lexical features, syntactic features and cue phrases, the proposed discourse parser constructs RS trees using high-level semantic features inherited from the Universal Networking Language (UNL). The UNL also adds a language-independent quality to the parser, because the UNL represents texts in a language-independent manner. The parser uses a naive Bayes probabilistic classifier to label discourse relations. It has been tested using 500 Tamil-language documents and the Rhetorical Structure Theory Discourse Treebank, which comprises 21 English-language documents. The performance of the naive Bayes classifier has been compared with that of the support vector machine (SVM) classifier, which has been used in the earlier approaches to build a discourse parser. It is seen that the naive Bayes probabilistic classifier is better suited for discourse relation labeling when compared with the SVM classifier, in terms of training time, testing time, and accuracy. [ABSTRACT FROM AUTHOR]

Recently deep learning has been successfully used in speech recognition, however it has not been carefully explored and widely accepted for speaker verification. To incorporate deep learning into speaker verification, this paper proposes novel approaches of extracting and using features from deep learning models for text-dependent speaker verification. In contrast to the traditional short-term spectral feature, such as MFCC or PLP, in this paper, outputs from hidden layer of various deep models are employed as deep features for text-dependent speaker verification. Fours types of deep models are investigated: deep Restricted Boltzmann Machines, speech-discriminant Deep Neural Network (DNN), speaker-discriminant DNN, and multi-task joint-learned DNN. Once deep features are extracted, they may be used within either the GMM-UBM framework or the identity vector (i-vector) framework. Joint linear discriminant analysis and probabilistic linear discriminant analysis are proposed as effective back-end classifiers for identity vector based deep features. These approaches were evaluated on the RSR2015 data corpus. Experiments showed that deep feature based methods can obtain significant performance improvements compared to the traditional baselines, no matter if they are directly applied in the GMM-UBM system or utilized as identity vectors. The EER of the best system using the proposed identity vector is 0.10%, only one fifteenth of that in the GMM-UBM baseline. [ABSTRACT FROM AUTHOR]

Boosting is a classification method which has been proven useful in non-satellite image processing while it is still new to satellite remote sensing. It is a meta-algorithm, which builds a strong classifier from many weak ones in iterative way. We adapt the AdaBoost.M1 boosting algorithm in a new land cover classification scenario based on utilization of very simple threshold classifiers employing spectral and contextual information. Thresholds for the classifiers are automatically calculated adaptively to data statistics. The proposed method is employed for the exemplary problem of artificial area identification. Classification of IKONOS multispectral data results in short computational time and overall accuracy of 94.4% comparing to 94.0% obtained by using AdaBoost.M1 with trees and 93.8% achieved using Random Forest. The influence of a manipulation of the final threshold of the strong classifier on classification results is reported. [ABSTRACT FROM AUTHOR]

In Japanese, numeral classifiers--or measure words--co-occur with numbers in counting phrases. The present study characterized parent numeral classifier use and its relation to children's classifier acquisition and number learning. Twenty-four Japanese-speaking parents and their two- to six-year-old children viewed and talked about two wordless picture books about counting to each other. Children also participated in a Counting task and Give-N task. Results revealed (1) parents' classifier use changed in relation to children's age and classifier use, and (2) parents' increased use of specific classifiers was uniquely associated with children's number understanding. These results suggest that aspects of children's language and numerical development are related to parents' language input, demonstrating the importance of examining the relation between language and cognition in a developmental context. [ABSTRACT FROM AUTHOR]

Validating online stream classifiers has traditionally assumed the availability of labeled samples, which can be monitored over time, to detect concept drift. However, labeling in streaming domains is expensive, time consuming and in certain applications, such as land mine detection, not a possibility at all. In this paper, the Margin Density Drift Detection (MD3) approach is proposed, which can signal change using unlabeled samples and requires labeling only for retraining, in the event of a drift. The MD3 approach when evaluated on 5 synthetic and 5 real world drifting data streams, produced statistically equivalent classification accuracy to that of a fully labeled accuracy tracking drift detector, and required only a third of the samples to be labeled, on average. [ABSTRACT FROM AUTHOR]

Trial and error classifiers, corresponding to concepts which change their extensions over time, are introduced and briefly philosophically motivated. A fragment of the language of classical first-order logic is given a new semantics, using $$\omega $$ -sequences of classical models, in order to interpret the basic predicates as classifiers of this kind. It turns out that we can use a natural deduction proof system which differs from classical logic only in the conditions for application of existential elimination. Soundness and completeness theorems are proved for this system. [ABSTRACT FROM AUTHOR]

We present a system for grasping unknown objects, even from piles or cluttered scenes, given a point cloud. Our method is based on the topography of a given scene and abstracts grasp-relevant structures to enable machine learning techniques for grasping tasks. We describe how Height Accumulated Features (HAF) and their extension, Symmetry Height Accumulated Features, extract grasp relevant local shapes. We investigate grasp quality using an F-score metric. We demonstrate the gain and the expressive power of HAF by comparing its trained classifier with one that resulted from training on simple height grids. An efficient way to calculate HAF is presented. We describe how the trained grasp classifier is used to explore the whole grasp space and introduce a heuristic to find the most robust grasp. We show how to use our approach to adapt the gripper opening width before grasping. In robotic experiments we demonstrate different aspects of our system on three robot platforms: a Schunk seven-degree-of-freedom arm, a PR2 and a Kuka LWR arm. We perform tasks to grasp single objects, autonomously unload a box and clear the table. Thereby we show that our approach is easily adaptable and robust with respect to different manipulators. As part of the experiments we compare our algorithm with a state-of-the-art method and show significant improvements. Concrete examples are used to illustrate the benefit of our approach compared with established grasp approaches. Finally, we show advantages of the symbiosis between our approach and object recognition. [ABSTRACT FROM AUTHOR]

Classical linear neural network architectures, such as the optimal linear associative memory (OLAM) Kohonen and Ruohonen (IEEE Trans Comp 22(7):701-702, ) and the adaptive linear element (Adaline) Widrow (IEEE Signal Process Mag 22(1):100-106, ; Widrow and Winter (IEEE Comp 21(3):25-39, ), are commonly used either as a standalone pattern classifier for linearly separable problems or as a fundamental building block of multilayer nonlinear classifiers, such as the multilayer perceptron (MLP), the radial basis functions networks (RBFN), the extreme learning machine (ELM) (Int J Mach Learn Cyber 2:107-122, ) and the echo-state network (ESN) Emmerich (Proceedings of the 20th international conference on artificial neural networks, 148-153, ). A common feature shared by the learning equations of OLAM and Adaline, respectively, the ordinary least squares (OLS) and the least mean squares (LMS) algorithms, is that they are optimal only under the assumption of gaussianity of the errors. However, the presence of outliers in the data causes the error distribution to depart from gaussianity and hence the classifier performance deteriorates. Bearing this in mind, in this paper we develop simple and efficient extensions of OLAM and Adaline, named Robust OLAM (ROLAM) and Robust Adaline (Radaline), which are robust to labeling errors (a.k.a. label noise), a type of outlier that often occur in classification tasks. This type of outlier usually results from mistakes during labelling the data points (e.g. misjudgement of a specialist) or from typing errors during creation of data files (e.g. by striking an incorrect key on a keyboard). To deal with such outliers, the ROLAM and the Radaline use $$M$$ -estimators to compute the weights of the OLAM and Adaline networks, instead of using standard OLS/LMS algorithms. By means of comprehensive computer simulations using synthetic and real-world data sets, we show that the proposed robust linear classifiers consistently outperforms their original versions. [ABSTRACT FROM AUTHOR]

Albeit automated classifiers offer a standard tool in many application areas, there exists hardly a generic possibility to directly inspect their behavior, which goes beyond the mere classification of (sets of) data points. In this contribution, we propose a general framework how to visualize a given classifier and its behavior as concerns a given data set in two dimensions. More specifically, we use modern nonlinear dimensionality reduction (DR) techniques to project a given set of data points and their relation to the classification decision boundaries. Furthermore, since data are usually intrinsically more than two-dimensional and hence cannot be projected to two dimensions without information loss, we propose to use discriminative DR methods which shape the projection according to given class labeling as is the case for a classification setting. With a given data set, this framework can be used to visualize any trained classifier which provides a probability or certainty of the classification together with the predicted class label. We demonstrate the suitability of the framework in the context of different dimensionality reduction techniques, in the context of different attention foci as concerns the visualization, and as concerns different classifiers which should be visualized. [ABSTRACT FROM AUTHOR]

This paper introduces a new classifier design method based on a modification of the classical fuzzy c -means clustering. First, a new fuzzy $c$-means clustering with $p$ constant prototypes is proposed. This method can be considered a generalization of the concept of the conditional fuzzy clustering with some prototypes a priori known. A special initialization of the prototypes is introduced. Next, the proposed clustering method is used to construct the premises of an if–then rule-based classifier. The conclusions of these rules are obtained by minimization of a criterion function with various approximations of a misclassification error (e.g., based on the quadratic, the linear, the sigmoidal or the Huber's loss function). The conjugate gradient algorithm is used to minimize the proposed criterion function. Each if–then rule is represented in the Mamdani–Assilan form, which has good interpretability. Finally, an extensive experimental analysis on 14 benchmark datasets is performed to demonstrate the validity of the classifier introduced. Its competitiveness to the state-of-the-art classifiers, with respect to both performance and interpretability, is also shown. [ABSTRACT FROM PUBLISHER]

One-class classification aims at constructing a distinctive classifier based on one class of examples. Most of the existing one-class classification methods are proposed based on the assumptions that: (1) there are a large number of training examples available for learning the classifier; (2) the training examples can be explicitly collected and hence do not contain any uncertain information. However, in real-world applications, these assumptions are not always satisfied. In this paper, we propose a novel approach called uncertain one-class transfer learning with support vector machine (UOCT-SVM), which is capable of constructing an accurate classifier on the target task by transferring knowledge from multiple source tasks whose data may contain uncertain information. In UOCT-SVM, the optimization function is formulated to deal with uncertain data and transfer learning based on one-class SVM. Then, an iterative framework is proposed to solve the optimization function. Extensive experiments have showed that UOCT-SVM can mitigate the effect of uncertain data on the decision boundary and transfer knowledge from source tasks to help build an accurate classifier on the target task, compared with state-of-the-art one-class classification methods. [ABSTRACT FROM AUTHOR]

In this article, we consider a geometric classifier that is applicable to multiclass classification for high-dimensional data. We show the consistency property and the asymptotic normality of the geometric classifier under certain mild conditions. We discuss sample size determination so that the geometric classifier can ensure that its misclassification rates are less than prespecified thresholds. We give a two-stage procedure to estimate the sample sizes required in such a geometric classifier and propose amisclassification rate–adjusted classifier (MRAC)based on the geometric classifier. We evaluate the performance of the MRAC theoretically and numerically. Finally, we demonstrate the MRAC in actual data analyses by using a microarray data set. [ABSTRACT FROM PUBLISHER]

Naive Bayes (NB) is one of the top 10 algorithms thanks to its simplicity, efficiency, and interpretability. To weaken its attribute independence assumption, naive Bayes tree (NBTree) has been proposed. NBTree is a hybrid algorithm, which deploys a naive Bayes classifier on each leaf node of the built decision tree and has demonstrated remarkable classification performance. When comes to text classification tasks, multinomial naive Bayes (MNB) has been a dominant modeling approach after the multi-variate Bernoulli model. Inspired by the success of NBTree, we propose a new algorithm called multinomial naive Bayes tree (MNBTree) by deploying a multinomial naive Bayes text classifier on each leaf node of the built decision tree. Different from NBTree, MNBTree builds a binary tree, in which the split attributes' values are just divided into zero and nonzero. At the same time, MNBTree uses the information gain measure instead of the classification accuracy measure to build the tree for reducing the time consumption. To further scale up the classification performance of MNBTree, we propose its multiclass learning version called multiclass multinomial naive Bayes tree (MMNBTree) by applying the multiclass technique to MNBTree. The experimental results on a large number of widely used text classification benchmark datasets validate the effectiveness of our proposed algorithms: MNBTree and MMNBTree. [ABSTRACT FROM AUTHOR]

This paper presents an unsupervised method for developing a character-based n-gram classifier that identifies loanwords or transliterated foreign words in Korean text. The classifier is trained on an unlabeled corpus using the Expectation Maximization algorithm, building on seed words extracted from the corpus. Words with high token frequency serve as native seed words. Words with seeming traces of vowel insertion to repair consonant clusters serve as foreign seed words. What counts as a trace of insertion is determined using phoneme co-occurrence statistics in conjunction with ideas and findings in phonology. Experiments show that the method can produce an unsupervised classifier that performs at a level comparable to that of a supervised classifier. In a cross-validation experiment using a corpus of about 9.2 million words and a lexicon of about 71,000 words, mean F-scores of the best unsupervised classifier and the corresponding supervised classifier were 94.77 and 96.67 %, respectively. Experiments also suggest that the method can be readily applied to other languages with similar phonotactics such as Japanese. [ABSTRACT FROM AUTHOR]

Least Squares Twin Support Vector Machine (LSTSVM) is a binary classifier and the extension of it to multiclass is still an ongoing research issue. In this paper, we extended the formulation of binary LSTSVM classifier to multi-class by using the concepts such as “One-versus-All”, “One-versus-One”, “All-versus-One” and Directed Acyclic Graph (DAG). This paper performs a comparative analysis of these multi-classifiers in terms of their advantages, disadvantages and computational complexity. The performance of all the four proposed classifiers has been validated on twelve benchmark datasets by using predictive accuracy and training–testing time. All the proposed multi-classifiers have shown better performance as compared to the typical multi-classifiers based on ‘Support Vector Machine’ and ‘Twin Support Vector Machine’. Friedman’s statistic and Nemenyi post hoc tests are also used to test significance of predictive accuracy differences between classifiers. [ABSTRACT FROM AUTHOR]

The use of 'Big Data' in policy and decision making is a current topic of debate. The 2013 murder of Drummer Lee Rigby in Woolwich, London, UK led to an extensive public reaction on social media, providing the opportunity to study the spread of online hate speech (cyber hate) on Twitter. Human annotated Twitter data was collected in the immediate aftermath of Rigby's murder to train and test a supervised machine learning text classifier that distinguishes between hateful and/or antagonistic responses with a focus on race, ethnicity, or religion; and more general responses. Classification features were derived from the content of each tweet, including grammatical dependencies between words to recognize 'othering' phrases, incitement to respond with antagonistic action, and claims of well-founded or justified discrimination against social groups. The results of the classifier were optimal using a combination of probabilistic, rule-based, and spatial-based classifiers with a voted ensemble meta-classifier. We demonstrate how the results of the classifier can be robustly utilized in a statistical model used to forecast the likely spread of cyber hate in a sample of Twitter data. The applications to policy and decision making are discussed. [ABSTRACT FROM AUTHOR]

In Mandarin Chinese, the phrasal modifier or complement of a noun, usually followed by the functional element de, may occur either between a classifier and the noun, or at the left edge of the whole nominal, preceding a (demonstrative)-numeral-classifier string. The latter order exhibits many restrictions and thus is marked. Some of the restrictions are similar to the restrictions on the marked nominal-internal order in English. This paper argues that the marked order in Chinese is derived by a phrasal movement. The movement analysis is supported by the hierarchy of nominal-internal elements, the Superiority Condition effect, the Crossing-Over-Nesting effect, and island effect. This movement analysis is able to explain certain restrictions on the marked order, with respect to the types of modifiers, the readings of the same modifier, and zheyang 'such' at the left edge. The analysis also provides a possible account for the absence of nonspecific readings and the impossibility of de-stranding for the marked order. [ABSTRACT FROM AUTHOR]

Real-world datasets often contain large numbers of unlabeled data points, because there is additional cost for obtaining the labels. Semi-supervised learning (SSL) algorithms use both labeled and unlabeled data points for training that can result in higher classification accuracy on these datasets. Generally, traditional SSLs tentatively label the unlabeled data points on the basis of the smoothness assumption that neighboring points should have the same label. When this assumption is violated, unlabeled points are mislabeled injecting noise into the final classifier. An alternative SSL approach is cluster-then-label (CTL), which partitions all the data points (labeled and unlabeled) into clusters and creates a classifier by using those clusters. CTL is based on the less restrictive cluster assumption that data points in the same cluster should have the same label. As shown, this allows CTLs to achieve higher classification accuracy on many datasets where the cluster assumption holds for the CTLs, but smoothness does not hold for the traditional SSLs. However, cluster configuration problems (e.g., irrelevant features, insufficient clusters, and incorrectly shaped clusters) could violate the cluster assumption. We propose a new framework for CTLs by using a genetic algorithm (GA) to evolve classifiers without the cluster configuration problems (e.g., the GA removes irrelevant attributes, updates number of clusters, and changes the shape of the clusters). We demonstrate that a CTL based on this framework achieves comparable or higher accuracy with both traditional SSLs and CTLs on 12 University of California, Irvine machine learning datasets. [ABSTRACT FROM AUTHOR]

This paper investigates whether individual speakers forming a homogeneous group differ in their choice and pronunciation of words when engaged in casual conversation, and if so, how they differ. More specifically, it examines whether the Balanced Winnow classifier is able to distinguish between the twenty speakers of the Ernestus Corpus of Spontaneous Dutch, who all have the same social background. To examine differences in choice and pronunciation of words, instead of characteristics of the speech signal itself, classification was based on lexical and pronunciation features extracted from hand-made orthographic and automatically generated broad phonetic transcriptions. The lexical features consisted of words and two-word combinations. The pronunciation features represented pronunciation variations at the word and phone level that are typical for casual speech. The best classifier achieved a performance of 79.9% and was based on the lexical features and on the pronunciation features representing single phones and triphones. The speakers must thus differ from each other in these features. Inspection of the relevant features indicated that, among other things, the words relevant for classification generally do not contain much semantic content, and that speakers differ not only from each other in the use of these words but also in their pronunciation. [ABSTRACT FROM AUTHOR]

In a classifier language like Chinese, all the noun phrases with numerals or demonstratives must go with classifiers, for example, liǎng zhī bǐ 兩枝筆 'two CL pen.' Classifiers vary depending on the nouns. This poses a big challenge for learners of Chinese as second language. Cognition-based multimedia and paper materials for Chinese material classifiers were developed and proved to be effective for learners of Chinese as a second language. The experimental group ( N = 15) used the multimedia program and the control group ( N= 20) received the same information on paper. Both groups made significant progress after a 10-week treatment. The results suggest cognitive principles are helpful for classifier learning in either multimedia or paper format. [ABSTRACT FROM AUTHOR]