270 results on '"Pechenizkiy, M."'
Search Results
2. CHBias: Bias Evaluation and Mitigation of Chinese Conversational Language Models
- Author
-
Zhao, J, Fang, M, Shi, Z, Li, Y, Chen, L, Pechenizkiy, M, Zhao, J, Fang, M, Shi, Z, Li, Y, Chen, L, and Pechenizkiy, M
- Abstract
Pretrained conversational agents have been exposed to safety issues, exhibiting a range of stereotypical human biases such as gender bias. However, there are still limited bias categories in current research, and most of them only focus on English. In this paper, we introduce a new Chinese dataset, CHBias, for bias evaluation and mitigation of Chinese conversational language models. Apart from those previous well-explored bias categories, CHBias includes under-explored bias categories, such as ageism and appearance biases, which received less attention. We evaluate two popular pretrained Chinese conversational models, CDial-GPT and EVA2.0, using CHBias. Furthermore, to mitigate different biases, we apply several debiasing methods to the Chinese pretrained models. Experimental results show that these Chinese pretrained models are potentially risky for generating texts that contain social biases, and debiasing methods using the proposed dataset can make response generation less biased while preserving the models' conversational capabilities. more...
- Published
- 2023
Catalog
3. A comparative study of dimensionality reduction techniques to enhance trace clustering performances
- Author
-
Song, M., Yang, H., Siadat, S.H., and Pechenizkiy, M.
- Published
- 2013
- Full Text
- View/download PDF
4. Mining sequences with exceptional transition behaviour of varying order using quality measures based on information-theoretic scoring functions
- Author
-
Schouten, R.M., De Paula Bueno, M.L., Duivesteijn, W., Pechenizkiy, M., Schouten, R.M., De Paula Bueno, M.L., Duivesteijn, W., and Pechenizkiy, M.
- Abstract
Contains fulltext : 285105.pdf (Publisher’s version ) (Open Access), Discrete Markov chains are frequently used to analyse transition behaviour in sequential data. Here, the transition probabilities can be estimated using varying order Markov chains, where order k specifies the length of the sequence history that is used to model these probabilities. Generally, such a model is fitted to the entire dataset, but in practice it is likely that some heterogeneity in the data exists and that some sequences would be better modelled with alternative parameter values, or with a Markov chain of a different order. We use the framework of Exceptional Model Mining (EMM) to discover these exceptionally behaving sequences. In particular, we propose an EMM model class that allows for discovering subgroups with transition behaviour of varying order. To that end, we propose three new quality measures based on information-theoretic scoring functions. Our findings from controlled experiments show that all three quality measures find exceptional transition behaviour of varying order and are reasonably sensitive. The quality measure based on Akaike’s Information Criterion is most robust for the number of observations. We furthermore add to existing work by seeking for subgroups of sequences, as opposite to subgroups of transitions. Since we use sequence-level descriptive attributes, we form subgroups of entire sequences, which is practically relevant in situations where you want to identify the originators of exceptional sequences, such as patients. We show this relevance by analysing sequences of blood glucose values of adult persons with diabetes type 2. In the experiments, we find subgroups of patients based on age and glycated haemoglobin (HbA1c), a measure known to correlate with average blood glucose values. Clinicians and domain experts confirmed the transition behaviour as estimated by the fitted Markov chain models. more...
- Published
- 2022
5. Phrase-level Textual Adversarial Attack with Label Preservation
- Author
-
Lei, Y., Cao, Y., Li, D., Tianyi Zhou, Fang, M., and Pechenizkiy, M.
- Subjects
FOS: Computer and information sciences ,Computer Science - Computation and Language ,Artificial Intelligence (cs.AI) ,Computer Science - Artificial Intelligence ,Computation and Language (cs.CL) - Abstract
Generating high-quality textual adversarial examples is critical for investigating the pitfalls of natural language processing (NLP) models and further promoting their robustness. Existing attacks are usually realized through word-level or sentence-level perturbations, which either limit the perturbation space or sacrifice fluency and textual quality, both affecting the attack effectiveness. In this paper, we propose Phrase-Level Textual Adversarial aTtack (PLAT) that generates adversarial samples through phrase-level perturbations. PLAT first extracts the vulnerable phrases as attack targets by a syntactic parser, and then perturbs them by a pre-trained blank-infilling model. Such flexible perturbation design substantially expands the search space for more effective attacks without introducing too many modifications, and meanwhile maintaining the textual fluency and grammaticality via contextualized generation using surrounding texts. Moreover, we develop a label-preservation filter leveraging the likelihoods of language models fine-tuned on each class, rather than textual similarity, to rule out those perturbations that potentially alter the original class label for humans. Extensive experiments and human evaluation demonstrate that PLAT has a superior attack effectiveness as well as a better label consistency than strong baselines., Comment: NAACL-HLT 2022 Findings (Long), 9 pages + 2 pages references + 8 pages appendix more...
- Published
- 2022
- Full Text
- View/download PDF
6. Bridging learning sciences, machine learning and affective computing for understanding cognition and affect in collaborative learning
- Author
-
Järvelä, S. (Sanna), Gašević, D. (Dragan), Seppänen, T. (Tapio), Pechenizkiy, M. (Mykola), Kirschner, P. A. (Paul A.), Järvelä, S. (Sanna), Gašević, D. (Dragan), Seppänen, T. (Tapio), Pechenizkiy, M. (Mykola), and Kirschner, P. A. (Paul A.) more...
- Abstract
Collaborative learning (CL) can be a powerful method for sharing understanding between learners. To this end, strategic regulation of processes, such as cognition and affect (including metacognition, emotion and motivation) is key. Decades of research on self‐regulated learning has advanced our understanding about the need for and complexity of those mediating processes in learning. Recent research has shown that it is not only the individual’s but also the group’s shared processes that matter and, thus, that regulation at the group level is critical for learning success. A problem here is that the “shared” processes in CL are invisible, which makes it almost impossible for researchers to study and understand them, for learners to recognize them and for teachers to support them. Traditionally, research has not been able to make these processes visible nor has it been able to collect data about them. With the aid of advanced technologies, signal processing and machine learning, we are on the verge of “seeing” these complex phenomena and understanding how they interact. We posit that technological solutions and digital tools available today and in the future will help advance the theory underlying the cognitive, metacognitive, emotional and social components of individual, peer and group learning when seen through a multidisciplinary lens. The aim of this paper is to discuss and demonstrate how multidisciplinary collaboration among the learning sciences, affective computing and machine learning is applied for understanding and facilitating CL. more...
- Published
- 2020
7. Bias Disparity in Collaborative Recommendation: Algorithmic Evaluation and Comparison
- Author
-
Mansoury, M., Mobasher, B., Robin Burke, and Pechenizkiy, M.
- Subjects
Social and Information Networks (cs.SI) ,FOS: Computer and information sciences ,Computer Science - Machine Learning ,Computer Science - Social and Information Networks ,Information Retrieval (cs.IR) ,Computer Science - Information Retrieval ,Machine Learning (cs.LG) - Abstract
Research on fairness in machine learning has been recently extended to recommender systems. One of the factors that may impact fairness is bias disparity, the degree to which a group's preferences on various item categories fail to be reflected in the recommendations they receive. In some cases biases in the original data may be amplified or reversed by the underlying recommendation algorithm. In this paper, we explore how different recommendation algorithms reflect the tradeoff between ranking quality and bias disparity. Our experiments include neighborhood-based, model-based, and trust-aware recommendation algorithms., Comment: Workshop on Recommendation in Multi-Stakeholder Environments (RMSE) at ACM RecSys 2019, Copenhagen, Denmark more...
- Published
- 2019
- Full Text
- View/download PDF
8. Multi-strategy differential evolution
- Author
-
Yaman, A., Iacca, G., Coler, M., Fletcher, G.H.L., Pechenizkiy, M., Sim, Kevin, Kaufmann, Paul, Integrated Circuits, Database Group, Data Mining, and Culture, Language & Technology
- Subjects
Continuous optimization ,Mathematical optimization ,education.field_of_study ,Optimization problem ,Computer science ,Parameter control ,Population ,020206 networking & telecommunications ,02 engineering and technology ,Evolutionary computation ,Public records ,Differential evolution ,Continuous optimization, Differential evolution, Parameter control, Strategy adaptation ,0202 electrical engineering, electronic engineering, information engineering ,Benchmark (computing) ,020201 artificial intelligence & image processing ,Strategy adaptation ,education ,Adaptation (computer science) ,Evolutionary Computation ,optimization problem - Abstract
We propose the Multi-strategy Differential Evolution (MsDE) algorithm to construct and maintain a self-adaptive ensemble of search strategies while solving an optimization problem. The ensemble of strategies is represented as agents that interact with the candidate solutions to improve their fitness. In the proposed algorithm, the performance of each agent is measured so that successful strategies are promoted within the ensemble. We propose two performance measures, and show their effectiveness in selecting successful strategies. We then present three population adaptation mechanisms, based on sampling, clone-best and clone-multiple adaptation schemes. The MsDE with different performance measures and population adaptation schemes is tested on the CEC2013 benchmark functions and compared with basic DE and with Self-Adaptive DE (SaDE). Our results show that MsDE is capable of efficiently adapting the strategies and parameters of DE and providing competitive results with respect to the state-of-the-art. more...
- Published
- 2018
- Full Text
- View/download PDF
9. Modeling brain responses to perceived speech with LSTM networks
- Author
-
Berezutskaya, Y., Freudenburg, Z.V., Ramsey, N.F., Güçlü, U., Gerven, M.A.J. van, Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Menkovski, V., Postma, E.J., Vanschoren, J., Putten, P. van der, Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Menkovski, V., Postma, E.J., Vanschoren, J., and Putten, P. van der more...
- Subjects
Language in Interaction ,Brain Networks and Neuronal Communication [DI-BCB_DCC_Theme 4] ,Cognitive artificial intelligence - Abstract
Item does not contain fulltext We used recurrent neural networks with longshort term memory units (LSTM) to model the brain responses to speech based on the speech audio features. We compared the performance of the LSTM models to the performance of the linear ridge regression model and found the LSTM models to be more robust for predicting brain responses across different feature sets. Benelearn 2017: Twenty-Sixth Benelux Conference on Machine Learning (Eindhoven, Netherlands, 9-10 June 2017) more...
- Published
- 2017
10. Automated decision-making fairness in an AI-driven world: Public perceptions, hopes and concerns
- Author
-
Araujo, T.B., Vreese, C.H. de, Helberger, N., Kruikemeier, S., Weert, J.C.M. van, Bol, N., Oberski, D., Pechenizkiy, M., Schaap, G.J., Taylor, L., Corporate Communication (ASCoR, FMG), Political Communication & Journalism (ASCoR, FMG), IViR (FdR), and Persuasive Communication (ASCoR, FMG) more...
- Subjects
Communication and Media - Abstract
Item does not contain fulltext Ongoing advances in artificial intelligence (AI) are increasingly part of scientific efforts as well as the public debate and the media agenda, raising hopes and concerns about the impact of automated decision making across different sectors of our society. This topic is receiving increasing attention at both national and cross- national levels. The present report contributes to informing this public debate, providing the results of a survey with 958 participants recruited from high-quality sample of the Dutch population. It provides an overview of public knowledge, perceptions, hopes and concerns about the adoption of AI and ADM across different societal sectors in the Netherlands. This report is part of a research collaboration between the Universities of Amsterdam, Tilburg, Radboud, Utrecht and Eindhoven (TU/e) on automated decision making, and forms input to the groups’ research on fairness in automated decision making. 21 p. more...
- Published
- 2018
11. Detection of alcoholism based on EEG signals and functional brain network features extraction
- Author
-
Ahmadi, N., Pei, Y., Pechenizkiy, M., Bamidis, Panagiotis D., Konstantinidis, Stathis Th., Rodrigues, Pedro Pereira, and Data Mining
- Subjects
Computer science ,Speech recognition ,Feature extraction ,02 engineering and technology ,Electroencephalography ,Brain Network ,SDG 3 – Goede gezondheid en welzijn ,Lateralization of brain function ,Feature Extraction ,03 medical and health sciences ,0302 clinical medicine ,Discriminative model ,SDG 3 - Good Health and Well-being ,0202 electrical engineering, electronic engineering, information engineering ,medicine ,EEG ,Entropy (energy dispersal) ,medicine.diagnostic_test ,business.industry ,Wavelet transform ,Pattern recognition ,Classification ,Random forest ,Support vector machine ,Alcoholism ,Brain Signal Processing ,020201 artificial intelligence & image processing ,Artificial intelligence ,business ,030217 neurology & neurosurgery - Abstract
Alcoholism is a common disorder that leads to brain defects and associated cognitive, emotional and behavioral impairments. Finding and extracting discriminative biological markers, which are correlated to healthy brain pattern and alcoholic brain pattern, helps us to utilize automatic methods for detecting and classifying alcoholism. Many brain disorders could be detected by analysing the Electroencephalography (EEG) signals. In this paper, for extracting the required markers we analyse the EEG signals for two groups of alcoholic and control subjects. Then by applying wavelet transform, band-limited EEG signals are decomposed into five frequency sub-bands. Also, the principle component analysis (PCA) is employed to choose the most information carrying channels. By examining various features from different frequency sub-bands, six discriminative features for classification are selected. From functional brain network perspective, the lower synchronization in Beta frequency sub-band and loss of lateralization in Alpha frequency sub-band in alcoholic subjects are observed. Also from signal processing perspective we found that alcoholic subjects have lower values of fractal dimension, energy and entropy compared to control ones. Five different classifiers are used to classify these groups of alcoholic and control subjects that show very high accuracies (more than 90%). However, by comparing the performance of different classifiers, SVM, random forest and gradient boosting show the best performances with accuracies near 100%. Our study shows that fractal dimension, entropy and energy of channel C1 in Alpha frequency sub-band are the more important features for classification. more...
- Published
- 2017
12. An in-situ trainable gesture classifier
- Author
-
van Diepen, A., Cox, M.G.H., de Vries, A., Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Signal Processing Systems, and Bayesian Intelligent Autonomous Systems Lab
- Subjects
InformationSystems_INFORMATIONINTERFACESANDPRESENTATION(e.g.,HCI) - Abstract
Gesture recognition, i.e., the recognition of pre-defined gestures by arm or hand movements, enables a natural extension of the way we currently interact with devices (Horsley, 2016). Commercially available gesture recognition systems are usually pre-trained: the developers specify a set of gestures, and the user is provided with an algorithm that can recognize just these gestures. To improve the user experience, it is often desirable to allow users to define their own gestures. In that case, the user needs to train the recognition system herself by a set of example gestures. Crucially, this scenario requires learning gestures from just a few training examples in order to avoid overburdening the user. We present a new in-situ trainable gesture classifier based on a hierarchical probabilistic modeling approach. Casting both learning and recognition as probabilistic inference tasks yields a principled way to design and evaluate algorithm candidates. Moreover, the Bayesian approach facilitates learning of prior knowledge about gestures, which leads to fewer needed examples for training new gestures. more...
- Published
- 2017
13. Local process models: pattern mining with process models
- Author
-
Tax, N., Sidorova, N., van der Aalst, W.M.P., Duivesteijn, W., Pechenizkiy, M., Fletcher, G., Menkovski, V., Postma, E., Vanschoren, J., van der Putten, P., Information Systems WSK&I, and Process Science more...
- Subjects
ComputingMilieux_LEGALASPECTSOFCOMPUTING - Published
- 2017
14. Predictive business process monitoring with LSTMs
- Author
-
Tax, N., Verenich, I., La Rosa, M., Dumas, M., Duivesteijn, W., Pechenizkiy, M., Fletcher, G., Menkovski, V., Postma, E., Vanschoren, J., van der Putten, P., Information Systems WSK&I, and Process Science more...
- Subjects
ComputingMilieux_LEGALASPECTSOFCOMPUTING - Published
- 2017
15. A Gaussian process mixture prior for hearing loss modeling
- Author
-
Cox, M.G.H., de Vries, A., Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Signal Processing Systems, and Bayesian Intelligent Autonomous Systems Lab
- Abstract
Machine learning approaches to hearing loss estimation can significantly reduce the number of required experiments, but require a good probabilistic hearing loss model. In this work we introduce such a model, obtained by fitting a mixture of Gaussian processes to a vast database containing audiometric records of around 85k people. The learned model can be used as a prior distribution for hearing loss, and can be conditioned on age and gender. Evaluation on a test set shows that our model outperforms an optimized Gaussian process model in terms of predictive accuracy. more...
- Published
- 2017
16. Hunting the unknown: White-box database leakage detection
- Author
-
Costante, E., Hartog, den, J.I., Petkovic, M., Etalle, S., Pechenizkiy, M., Atluri, V., Pernul, G., Mathematics and Computer Science, Security, and Data Mining
- Subjects
Measure (data warehouse) ,Engineering ,business.industry ,Data security ,computer.software_genre ,Orders of magnitude (bit rate) ,Anomaly detection ,False positive rate ,Data mining ,Leakage (economics) ,Detection rate ,business ,Database transaction ,computer - Abstract
Data leakage causes significant losses and privacy breaches worldwide. In this paper we present a white-box data leakage detection system to spot anomalies in database transactions. We argue that our approach represents a major leap forward w.r.t. previous work because: i) it significantly decreases the False Positive Rate (FPR) while keeping the Detection Rate (DR) high; on our experimental dataset, consisting of millions of real enterprise transactions, we measure a FPR that is orders of magnitude lower than in state-of-the-art comparable approaches; and ii) the white-box approach allows the creation of self-explanatory and easy to update profiles able to explain why a given query is anomalous, which further boosts the practical applicability of the system. more...
- Published
- 2014
- Full Text
- View/download PDF
17. Finding predictive EEG complexity features for classification of epileptic and psychogenic nonepileptic seizures using imperialist competitive algorithm
- Author
-
Ahmadi, N., Carrette, Evelien, Aldenkamp, A.P., Pechenizkiy, M., Ahmadi, N., Carrette, Evelien, Aldenkamp, A.P., and Pechenizkiy, M.
- Abstract
In this study, the imperialist competitive algorithm (ICA) is applied for classification of epileptic seizure and psychogenic nonepileptic seizure (PNES). For this purpose, after decomposing the EEG signal into five sub-bands and extracting some complexity features of EEG, the ICA is applied to find the predictive feature subset that maximizes the classification performance in the frequency spectrum. Results show that the spectral entropy and Renyi entropy are the most important EEG features as they are always appeared in the best feature subsets when applying different classifiers. Also, it is observed that the SVM-RBF and SVM-linear models are the best classifiers resulting in highest performance metrics compared to other classifiers. Our study shows that the reported algorithm is able to classify the epileptic seizure and PNES with a very high classification metrics. more...
- Published
- 2018
18. ELBA: Exceptional Learning Behavior Analysis
- Author
-
Du, X., Duivesteijn, W., Klabbers, M.D., Pechenizkiy, M., Du, X., Duivesteijn, W., Klabbers, M.D., and Pechenizkiy, M.
- Abstract
Behavioral records collected through course assessments, peer assignments, and programming assignments in Massive Open Online Courses (MOOCs) provide multiple views about a student’s study style. Study behavior is correlated with whether or not the student can get a certificate or drop out from a course. It is of predominant importance to identify the particular behavioral patterns and establish an accurate predictive model for the learning results, so that tutors can give well-focused assistance and guidance on specific students. However, the behavioral records of individuals are usually very sparse; behavioral records between individuals are inconsistent in time and skewed in contents. These remain big challenges for the state-of-the-art methods. In this paper, we engage the concept of subgroup as a trade-off to overcome the sparsity of individual behavioral records and inconsistency between individuals. We employ the framework of Exceptional Model Mining (EMM) to discover exceptional student behavior. Various model classes of EMM are applied on dropout rate analysis, correlation analysis between length of learning behavior sequence and course grades, and passing state prediction analysis. Qualitative and quantitative experimental results on real MOOCs datasets show that our method can discover significantly interesting learning behavioral patterns of students. more...
- Published
- 2018
19. How to capitalise on mobility, proximity and motion analytics to support formal and informal education?
- Author
-
Martinez-Maldonado, R., Vanessa Echeverria, Yacef, K., Dos Santos, A. D. P., Pechenizkiy, M., and Data Mining
- Subjects
Mobility ,Motor learning ,Sensors ,Wearables ,Physical spaces ,Indoor localisation - Abstract
© 2017, CEUR-WS. All rights reserved. Learning Analytics and similar data-intensive approaches aimed at understanding and/or supporting learning have mostly focused on the analysis of students' data automatically captured by personal computers or, more recently, mobile devices. Thus, most student behavioural data are limited to the interactions between students and particular learning applications. However, learning can also occur beyond these interface interactions, for instance while students interact face-to-face with other students or their teachers. Alternatively, some learning tasks may require students to interact with non-digital physical tools, to use the physical space, or to learn in different ways that cannot be mediated by traditional user interfaces (e.g. motor and/or audio learning). The key questions here are: why are we neglecting these kinds of learning activities? How can we provide automated support or feedback to students during these activities? Can we find useful patterns of activity in these physical settings as we have been doing with computer-mediated settings? This position paper is aimed at motivating discussion through a series of questions that can justify the importance of designing technological innovations for physical learning settings where mobility, proximity and motion are tracked, just as digital interactions have been so far. more...
- Published
- 2017
20. BoostEMM : Transparent boosting using exceptional model mining
- Author
-
van der Zon, S.B., Zeev Ben Mordehay, O., Vrijdag, T.S., van Ipenburg, W., Veldsink, J., Duivesteijn, W., Pechenizkiy, M., Bordino, I., Caldarelli, G., Fumarola, F., Gullo, F., Squartini, T., Information Systems WSK&I, Mathematics and Computer Science, and Data Mining more...
- Subjects
ComputingMethodologies_PATTERNRECOGNITION ,Class imbalance ,Responsible analytics ,Exceptional Model Mining ,Model transparency ,Boosting - Abstract
Boosting is an iterative ensemble-learning paradigm. Every iteration, a weak predictor learns a classification task, taking into account performance achieved in previous iterations. This is done by assigning weights to individual records of the dataset, which are increased if the record is misclassified by the previous weak predictor. Hence, subsequent predictors learn to focus on problematic records in the dataset. Boosting ensembles such as AdaBoost have shown to be effective models at fighting both high variance and high bias, even in challenging situations such as class imbalance. However, some aspects of AdaBoost might imply limitations for its deployment in the real world. On the one hand, focusing on problematic records can lead to overfitting in the presence of random noise. On the other hand, learning a boosting ensemble that assigns higher weights to hard-to-classify people might throw up serious questions in the age of responsible and transparent data analytics; if a bank must tell a customer that they are denied a loan, because the underlying algorithm made a decision specifically focusing the customer since they are hard to classify, this could be legally dubious. To kill these two birds with one stone, we introduce BoostEMM: a variant of AdaBoost where in every iteration of the procedure, rather than boosting problematic records, we boost problematic subgroups as found through Exceptional Model Mining. Boosted records being part of a coherent group should prevent overfitting, and explicit definitions of the subgroups of people being boosted enhances the transparency of the algorithm. more...
- Published
- 2017
21. Techniques for discrimination-free predictive models
- Author
-
Kamiran, F., Calders, T.G.K., Pechenizkiy, M., Custers, B.H.M., Schermer, B.W., Zarsky, T.Z., and Data Mining
- Subjects
Engineering ,Gender discrimination ,SDG 5 - Gender Equality ,business.industry ,Decision tree learning ,media_common.quotation_subject ,Wage ,SDG 5 – Gendergelijkheid ,Space (commercial competition) ,computer.software_genre ,Machine learning ,Uncorrelated ,Task (project management) ,Constraint (information theory) ,Artificial intelligence ,Data mining ,business ,Data objects ,computer ,media_common - Abstract
In this chapter, we give an overview of the techniques developed ourselves for constructing discrimination-free classifiers. In discrimination-free classification the goal is to learn a predictive model that classifies future data objects as accurately as possible, yet the predicted labels should be uncorrelated to a given sensitive attribute. For example, the task could be to learn a gender-neutral model that predicts whether a potential client of a bank has a high income or not. The techniques we developed for discrimination-aware classification can be divided into three categories: (1) removing the discrimination directly from the historical dataset before an off-the-shelf classification technique is applied; (2) changing the learning procedures themselves by restricting the search space to non-discriminatory models; and (3) adjusting the discriminatory models, learnt by off-the-shelf classifiers on discriminatory historical data, in a post-processing phase. Experiments show that even with such a strong constraint as discrimination-freeness, still very accurate models can be learnt. In particular,we study a case of income prediction,where the available historical data exhibits a wage gap between the genders. Due to legal restrictions, however, our predictions should be gender-neutral. The discrimination-aware techniques succeed in significantly reducing gender discrimination without impairing too much the accuracy. more...
- Published
- 2013
- Full Text
- View/download PDF
22. Context-aware personal route recognition
- Author
-
Mazhelis, O., Zliobaite, I., Pechenizkiy, M., Elomaa, T., Hollmén, J., Mannila, H., and Data Mining
- Subjects
Point (typography) ,business.industry ,Computer science ,Context (language use) ,Recommender system ,computer.software_genre ,Computer security ,Machine learning ,Trajectory ,Route planning software ,Artificial intelligence ,business ,Intelligent transportation system ,computer ,Personally identifiable information ,Drawback - Abstract
Personal route recognition is an important element of intelligent transportation systems. The results may be used for providing personal information about location-specific events, services, emergency or disaster situations, for location-specific advertising and more. Existing real-time route recognition systems often compare the current driving trajectory against the trajectories observed in past and select the most similar route as the most likely. The problem is that such systems are inaccurate in the beginning of a trip, as typically several different routes start at the same departure point (e.g. home). In such situations the beginnings of trajectories overlap and the trajectory alone is insufficient to recognize the route. This drawback limits the utilization of route prediction systems, since accurate predictions are needed as early as possible, not at the end of the trip. To solve this problem we incorporate external contextual information (e.g. time of the day) into route recognition from trajectory. We develop a technique to determine from the historical data how the probability of a route depends on contextual features and adjust (post-correct) the route recognition output accordingly. We evaluate the proposed context-aware route recognition approach using the data on driving behavior of twenty persons residing in Aalborg, Denmark, monitored over two months. The results confirm that utilizing contextual information in the proposed way improves the accuracy of route recognition, especially in cases when the historical routes highly overlap. more...
- Published
- 2011
- Full Text
- View/download PDF
23. Defining adaptation in a generic multi layer model: CAM: The GRAPPLE Conceptual Adaptation Model
- Author
-
Hendrix, M., De Bra, P.M.E., Pechenizkiy, M., Smits, D., Cristea, A.I., Dillenbourg, P., Specht, M., and Data Mining
- Subjects
Human–computer interaction ,business.industry ,Computer science ,Usability ,Adaptive hypermedia ,business ,Adaptation (computer science) ,Reference model ,Multi layer ,Graphical tools ,Task (project management) - Abstract
Authoring of Adaptive Hypermedia is a difficult and time consuming task. Reference models like LAOS and AHAM separate adaptation and content in different layers. Systems like AHA!, offer graphical tools based on these models to allow authors to define adaptation without knowing any adaptation language. The adaptation that can be defined using such tools is still limited. Authoring systems like MOT are more flexible, but usability of adaptation specification is low. This paper proposes a more generic model, CAM, which allows the adaptation to be defined in an arbitrary number of layers, where adaptation is expressed in terms of relationships between concepts. This model allows the creation of more powerful yet easier to use graphical authoring tools. more...
- Published
- 2008
- Full Text
- View/download PDF
24. Finding incident-related social media messages for emergency awareness
- Author
-
Nieuwenhuijse, A., Bakker, J., Pechenizkiy, M., Berendt, B., Bringmann, B., Fromont, E., Garriga, G., Miettinen, P., Tatti, N., Tresp, V., Information Systems WSK&I, and Data Mining
- Subjects
Information retrieval ,Emergency response ,SIMPLE (military communications protocol) ,Interface (Java) ,Computer science ,Relevance (information retrieval) ,Social media ,Decision-making ,Computer security ,computer.software_genre ,computer - Abstract
An information retrieval framework is proposed which searches for incident-related social media messages in an automated fashion. Using P2000 messages as an input for this framework and by extracting location information from text, using simple natural language processing techniques, a search for incident-related messages is conducted. A machine learned ranker is trained to create an ordering of the retrieved messages, based on their relevance. This provides an easy accessible interface for emergency response managers to aid them in their decision making process. more...
- Published
- 2016
25. Adaptive web-based educational application for autistic students
- Author
-
Montes Garcia, A., Stash, N., Fabri, M., De Bra, P.M.E., Fletcher, G.H.L., Pechenizkiy, M., Koidl, K., Steichen, Ben, Database Group, and Data Mining
- Subjects
ComputingMilieux_LEGALASPECTSOFCOMPUTING - Published
- 2016
26. Dynamic integration with random forests
- Author
-
Tsymbal, A., Pechenizkiy, M., Cunningham, P., Fürnkranz, J., Scheffer, T., Spiliopoulou, M., and Data Mining
- Subjects
Majority rule ,Similarity (geometry) ,business.industry ,Context (language use) ,Function (mathematics) ,Machine learning ,computer.software_genre ,Similitude ,Random forest ,Tree (data structure) ,Metric (mathematics) ,Artificial intelligence ,business ,computer ,Algorithm ,Mathematics - Abstract
Random Forests (RF) are a successful ensemble prediction technique that uses majority voting or averaging as a combination function. However, it is clear that each tree in a random forest may have a different contribution in processing a certain instance. In this paper, we demonstrate that the prediction performance of RF may still be improved in some domains by replacing the combination function with dynamic integration, which is based on local performance estimates. Our experiments also demonstrate that the RF Intrinsic Similarity is better than the commonly used Heterogeneous Euclidean/Overlap Metric in finding a neighbourhood for local estimates in the context of dynamic integration of classification random forests. more...
- Published
- 2006
- Full Text
- View/download PDF
27. Modeling brain responses to perceived speech with LSTM networks
- Author
-
Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Menkovski, V., Postma, E.J., Vanschoren, J., Putten, P. van der, Berezutskaya, Y., Freudenburg, Z.V., Ramsey, N.F., Güçlü, U., Gerven, M.A.J. van, Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Menkovski, V., Postma, E.J., Vanschoren, J., Putten, P. van der, Berezutskaya, Y., Freudenburg, Z.V., Ramsey, N.F., Güçlü, U., and Gerven, M.A.J. van more...
- Abstract
Benelearn 2017: Twenty-Sixth Benelux Conference on Machine Learning (Eindhoven, Netherlands, 9-10 June 2017), Item does not contain fulltext, We used recurrent neural networks with longshort term memory units (LSTM) to model the brain responses to speech based on the speech audio features. We compared the performance of the LSTM models to the performance of the linear ridge regression model and found the LSTM models to be more robust for predicting brain responses across different feature sets. more...
- Published
- 2017
28. Clustering-structure representative sampling from graph streams
- Author
-
Zhu, K., Pei, Y., Fletcher, G.H.L., Pechenizkiy, M., Zhang, J., Zhu, K., Pei, Y., Fletcher, G.H.L., Pechenizkiy, M., and Zhang, J.
- Abstract
Most existing sampling algorithms on graphs (i.e., network-structured data) focus on sampling from memory-resident static graphs and assume the entire graphs are always available. However, the graphs encountered in modern applications are often too large and/or too dynamic to be processed with limited memory. Furthermore, existing sampling techniques are inadequate for preserving the inherent clustering structure, which is an essential property of complex networks. To tackle these problems, we propose a new sampling algorithm that dynamically maintains a representative sample and is capable of retaining clustering structure in graph streams at any time. Performance of the proposed algorithm is evaluated through empirical experiments using real-world networks. The experimental results have shown that our proposed \textit{CPIES} algorithm can produce clustering-structure representative samples and outperforms current online sampling algorithms. more...
- Published
- 2017
29. Benelearn 2017: Proceedings of the Twenty-Sixth Benelux Conference on Machine Learning, Technische Universiteit Eindhoven, 9-10 June 2017
- Author
-
Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Menkovski, V., Postma, E.J., Vanschoren, J., van der Putten, P., Duivesteijn, W., Pechenizkiy, M., Fletcher, G.H.L., Menkovski, V., Postma, E.J., Vanschoren, J., and van der Putten, P. more...
- Published
- 2017
30. Clustering-structure representative sampling from graph streams
- Author
-
Zhang, J., Zhu, Kaijie, Pei, Y., Fletcher, G.H.L., Pechenizkiy, M., Zhang, J., Zhu, Kaijie, Pei, Y., Fletcher, G.H.L., and Pechenizkiy, M.
- Abstract
Most existing sampling algorithms on graphs (i.e., network-structured data) focus on sampling from memory-resident static graphs and assume the entire graphs are always available. However, the graphs encountered in modern applications are often too large and/or too dynamic to be processed with limited memory. Furthermore, existing sampling techniques are inadequate for preserving the inherent clustering structure, which is an essential property of complex networks. To tackle these problems, we propose a new sampling algorithm that dynamically maintains a representative sample and is capable of retaining clustering structure in graph streams at any time. Performance of the proposed algorithm is evaluated through empirical experiments using real-world networks. The experimental results have shown that our proposed \textit{CPIES} algorithm can produce clustering-structure representative samples and outperforms current online sampling algorithms. more...
- Published
- 2017
31. The nutcracker framework for ensemble interpretability
- Author
-
Zeev Ben Mordehay, O., Duivesteijn, W., Pechenizkiy, M., Zeev Ben Mordehay, O., Duivesteijn, W., and Pechenizkiy, M.
- Abstract
The basic principles behind ensembles (e.g. Random Forest, AdaBoost) are simple. But we’re still in trouble when attempting to explain the logic taken. Where does the problem lie? The reason that ensembles are effective is that the base estimators "work together" and compensate each for the others’ shortcomings. The Nutcracker Framework Given a trained ensemble and the relevant training / test dataset, construct prediction matrix, M, cases (rows) against predictions (columns). Bicluster M to a given number of R x C biclusters. Now, investigate performance per bicluster (R x C). Identify feature importance per base estimators group (C). Describe each of the R cases subgroups in terms of features and values. We use Exceptional Model Mining for that task. Performance of the ensemble against the dataset compared to performance of base estimator groups against subgroups of cases, adds transparency. more...
- Published
- 2017
32. BoostEMM: Transparent boosting using exceptional model mining
- Author
-
van der Zon, Simon, Mordehai, O. Zeev Ben, Vrijdag, T.S., van Ipenburg, W., Veldsink, J.W., Duivesteijn, W., Pechenizkiy, M., van der Zon, Simon, Mordehai, O. Zeev Ben, Vrijdag, T.S., van Ipenburg, W., Veldsink, J.W., Duivesteijn, W., and Pechenizkiy, M. more...
- Abstract
Boosting is an iterative ensemble-learning paradigm. Every iteration, a weak predictor learns a classification task, taking into account performance achieved in previous iterations. This is done by assigning weights to individual records of the dataset, which are increased if the record is misclassified by the previous weak predictor. Hence, subsequent predictors learn to focus on problematic records in the dataset. Boosting ensembles such as AdaBoost have shown to be effective models at fighting both high variance and high bias, even in challenging situations such as class imbalance. However, some aspects of AdaBoost might imply limitations for its deployment in the real world. On the one hand, focusing on problematic records can lead to overfitting in the presence of random noise. On the other hand, learning a boosting ensemble that assigns higher weights to hard-to-classify people might throw up serious questions in the age of responsible and transparent data analytics; if a bank must tell a customer that they are denied a loan, because the underlying algorithm made a decision specifically focusing the customer since they are hard to classify, this could be legally dubious. To kill these two birds with one stone, we introduce BoostEMM: a variant of AdaBoost where in every iteration of the procedure, rather than boosting problematic records, we boost problematic subgroups as found through Exceptional Model Mining. Boosted records being part of a coherent group should prevent overfitting, and explicit definitions of the subgroups of people being boosted enhances the transparency of the algorithm. more...
- Published
- 2017
33. Towards Proximity Tracking and Sensemaking for Supporting Teamwork and Learning
- Author
-
Martinez-Maldonado, R, Yacef, K, Santos, ADPD, Shum, SB, Echeverria, V, Santos, OC, Pechenizkiy, M, Martinez-Maldonado, R, Yacef, K, Santos, ADPD, Shum, SB, Echeverria, V, Santos, OC, and Pechenizkiy, M more...
- Abstract
© 2017 IEEE. A large number of learning tools offering some sort of personalisation features rely mainly on the analysis of logged interactions between students and particular user interfaces. Much less attention has been given to the analysis of physical aspects so often present in 'traditional' intellectual tasks, although these are both important in the full development of a life-long learner. This paper (1) discusses existing literature focused on supporting learning using proximity and location analytics and sensors, and, based on this, (2) illustrates the feasibility and potential of these analytics for teaching and learning through an study in the context of proximity and location analytics in a team-based health simulation classroom. more...
- Published
- 2017
34. How to capitalise on mobility, proximity and motion analytics to support formal and informal education?
- Author
-
Martinez-Maldonado, R, Echeverria, V, Yacef, K, Dos Santos, ADP, Pechenizkiy, M, Martinez-Maldonado, R, Echeverria, V, Yacef, K, Dos Santos, ADP, and Pechenizkiy, M
- Abstract
© 2017, CEUR-WS. All rights reserved. Learning Analytics and similar data-intensive approaches aimed at understanding and/or supporting learning have mostly focused on the analysis of students' data automatically captured by personal computers or, more recently, mobile devices. Thus, most student behavioural data are limited to the interactions between students and particular learning applications. However, learning can also occur beyond these interface interactions, for instance while students interact face-to-face with other students or their teachers. Alternatively, some learning tasks may require students to interact with non-digital physical tools, to use the physical space, or to learn in different ways that cannot be mediated by traditional user interfaces (e.g. motor and/or audio learning). The key questions here are: why are we neglecting these kinds of learning activities? How can we provide automated support or feedback to students during these activities? Can we find useful patterns of activity in these physical settings as we have been doing with computer-mediated settings? This position paper is aimed at motivating discussion through a series of questions that can justify the importance of designing technological innovations for physical learning settings where mobility, proximity and motion are tracked, just as digital interactions have been so far. more...
- Published
- 2017
35. Modelling embodied mobility teamwork strategies in a simulation-based healthcare classroom
- Author
-
Martinez-Maldonado, R, Buckingham-Shum, S, Pechenizkiy, M, Power, T, Hayes, C, Axisa, C, Martinez-Maldonado, R, Buckingham-Shum, S, Pechenizkiy, M, Power, T, Hayes, C, and Axisa, C
- Abstract
©2017 ACM. In many situations, it remains critical for team members to develop strategies to effectively use the space and tools available to complete demanding tasks. However, despite the availability of sensors and analytics for instrumenting physical space, relatively little progress has been made in modelling the embodied dimensions of co-located teamwork. This paper explores an in-The-wild pilot study through which we explore a methodology to model embodied mobility teamwork strategies in the context of healthcare education. We developed the means for tracking, clustering and processing student-nurses' mobility data around a patient manikin. We illustrate the feasibility of our approach by discussing ways to make sense of these data to uncover meaningful trends, and the inherent challenges of applying physical space analytics in authentic settings. more...
- Published
- 2017
36. The impact of feature extraction on the performance of a classifier : kNN, Naïve Bayes and C4.5
- Author
-
Pechenizkiy, M., Kégl, B., and Lapalme, G.
- Subjects
Covariance matrix ,Computer science ,business.industry ,Random projection ,Dimensionality reduction ,Feature extraction ,Linear classifier ,Pattern recognition ,Machine learning ,computer.software_genre ,Naive Bayes classifier ,ComputingMethodologies_PATTERNRECOGNITION ,Principal component analysis ,Artificial intelligence ,business ,computer ,Curse of dimensionality - Abstract
"The curse of dimensionality" is pertinent to many learning algorithms, and it denotes the drastic raise of computational complexity and the classification error in high dimensions. In this paper, different feature extraction techniques as means of (1) dimensionality reduction, and (2) constructive induction are analyzed with respect to the performance of a classifier. Three commonly used classifiers are taken for the analysis: kNN, Naïve Bayes and C4.5 decision tree. One of the main goals of this paper is to show the importance of the use of class information in feature extraction for classification and (in)appropriateness of random projection or conventional PCA to feature extraction for classification for some data sets. Two eigenvector-based approaches that take into account the class information are analyzed. The first approach is parametric and optimizes the ratio of between-class variance to the within-class variance of the transformed data. The second approach is a nonparametric modification of the first one based on the local calculation of the between-class covariance matrix. In experiments on benchmark data sets these two approaches are compared with each other, with conventional PCA, with random projection and with plain classification without feature extraction for each classifier. more...
- Published
- 2005
- Full Text
- View/download PDF
37. Feature extraction for classification in knowledge discovery systems
- Author
-
Pechenizkiy, M., Puuronen, S., Tsymbal, A., Palade, V., Howlett, R.J., and Jain, L.C.
- Subjects
Decision support system ,business.industry ,Computer science ,Dimensionality reduction ,Feature extraction ,Machine learning ,computer.software_genre ,Knowledge acquisition ,k-nearest neighbors algorithm ,Knowledge extraction ,Feature (computer vision) ,Artificial intelligence ,Data mining ,business ,computer ,Curse of dimensionality - Abstract
Dimensionality reduction is a very important step in the data mining process. In this paper, we consider feature extraction for classification tasks as a technique to overcome problems occurring because of "the curse of dimensionality". We consider three different eigenvector-based feature extraction approaches for classification. The summary of obtained results concerning the accuracy of classification schemes is presented and the issue of search for the most appropriate feature extraction method for a given data set is considered. A decision support system to aid in the integration of the feature extraction and classification processes is proposed. The goals and requirements set for the decision support system and its basic structure are defined. The means of knowledge acquisition needed to build up the proposed system are considered. more...
- Published
- 2003
- Full Text
- View/download PDF
38. Pattern-based emotion classification on social media
- Author
-
Tromp, E., Pechenizkiy, M., Medhat Gaber, M., Cocea, M., Wiratunga, N., Goker, A., and Data Mining
- Subjects
Computer science ,business.industry ,Polarity (physics) ,Emotion classification ,Emotion detection ,Sentiment analysis ,computer.software_genre ,Social media ,Granularity ,Artificial intelligence ,business ,Set (psychology) ,computer ,GeneralLiterature_REFERENCE(e.g.,dictionaries,encyclopedias,glossaries) ,Natural language processing - Abstract
Sentiment analysis can go beyond the typical granularity of polarity that assumes each text to be positive, negative or neural. Indeed, human emotions are much more diverse, and it is interesting to study how to define a more complete set of emotions and how to deduce these emotions from human-written messages. In this book chapter we argue that using Plutchik’s wheel of emotions model and a rule-based approach for emotion detection in text makes it a good framework for emotion classification on social media. We provide a detailed description of how to define rule-based patterns for Plutchik’s wheel emotion detection, how to learn them from the annotated social media and how to apply them for classifying emotions in the previously unseen texts. The results of the experimental study suggest that the described framework is promising and that it advances the current state-of-the-art in emotion detection. more...
- Published
- 2015
- Full Text
- View/download PDF
39. Matching and Maximizing? A neurally plausible model of stochastic reinforcement learning
- Author
-
Vroon, J.H., Rooij, I.J.E.I. van, Sprinkhuizen-Kuyper, I.G., Calders, T., Tuyls, K., Pechenizkiy, M., Calders, T., Tuyls, K., and Pechenizkiy, M.
- Subjects
Brain Networks and Neuronal Communication [DI-BCB_DCC_Theme 4] ,Perception, Action and Control [DI-BCB_DCC_Theme 2] ,Cognitive artificial intelligence ,GeneralLiterature_REFERENCE(e.g.,dictionaries,encyclopedias,glossaries) - Abstract
Contains fulltext : 77434.pdf (Publisher’s version ) (Open Access)
- Published
- 2009
40. Meaningful Representations Prevent Catastrophic Interference
- Author
-
Bieger, J., Sprinkhuizen-Kuyper, I.G., Rooij, I.J.E.I. van, Calders, T., Tuyls, K., Pechenizkiy, M., Calders, T., Tuyls, K., and Pechenizkiy, M.
- Subjects
Brain Networks and Neuronal Communication [DI-BCB_DCC_Theme 4] ,Computer Science::Neural and Evolutionary Computation ,Perception, Action and Control [DI-BCB_DCC_Theme 2] ,Cognitive artificial intelligence - Abstract
Contains fulltext : 77446.pdf (Publisher’s version ) (Open Access) Artificial Neural Networks (ANNs) attempt to mimic human neural networks in order to perform tasks. In order to do this, tasks need to be represented in ways that the network understands. In ANNs these representations are often arbitrary, whereas in humans it seems that these representations are often meaningful. This article shows how using more meaningful representations in ANNs can be very beneficial. We demonstrate that by using our Static Meaningful Representation Learning (SMRL) technique, ANNs can avoid the problem of catastrophic interference when sequentially learning multiple simple tasks. We also discuss how our approach overcomes known limitations of other techniques for dealing with catastrophic interference. more...
- Published
- 2009
41. Explaining soccer match outcomes with goal scoring opportunities predictive analytics
- Author
-
Eggels, H., van Elk, R., Pechenizkiy, M., Eggels, H., van Elk, R., and Pechenizkiy, M.
- Abstract
In elite soccer, decisions are often based on recent results and emotions. In this paper, we propose a method to determine the expected winner of a match in elite soccer. The expected result of a soccer match is determined by estimating the probability of scoring for the individual goal scoring opportunities. The outcome of a match is then obtained by integrating these probabilities. In our experimental study, we show that the probabilities of goal scoring opportunities accurately match reality. more...
- Published
- 2016
42. Editorial: a message from the editorial team and an introduction to the January-March 2016 issue
- Author
-
Brusilovsky, P., Sharples, M., Alves, G.R., Barnes, T., Chen, S.Y., Chu, C.H.C., Drachsler, H., Isotani, S., Lindsay, E., Ochoa, X., Pechenizkiy, M., Rodrigo, M.M.T., Romero, C., Sosnovsky, S., Ternier, S., Verbert, K., Brusilovsky, P., Sharples, M., Alves, G.R., Barnes, T., Chen, S.Y., Chu, C.H.C., Drachsler, H., Isotani, S., Lindsay, E., Ochoa, X., Pechenizkiy, M., Rodrigo, M.M.T., Romero, C., Sosnovsky, S., Ternier, S., and Verbert, K. more...
- Published
- 2016
43. WiBAF into a CMS: Personalization in learning environments made easy
- Author
-
García, AM, Stash, N, Fabri, M, De Bra, P, Fletcher, GHL, Pechenizkiy, M, García, AM, Stash, N, Fabri, M, De Bra, P, Fletcher, GHL, and Pechenizkiy, M
- Abstract
Adaptivity has proven successful in reducing navigation and comprehension problems in hypermedia documents. Authoring of adaptive hypermedia documents and especially of the adaptivity in these documents has been problematic or at least labour intensive throughout AH history. This paper shows how the integration of a CMS with an adaptive framework greatly simplifies the inclusion of personalization in existing educational applications. It does this within the context of European project Autism&Uni that uses adaptive hypermedia to offer information for students transitioning from high school to university, especially to cater for students on the autism spectrum as well as for non-autistic students. The use of our Within Browser adaptation framework (WiBAF) reduces privacy concerns because the user model is stored on the end-user's machine, and eliminates performance issues that currently prevent the adoption of adaptivity in MOOC platforms by having the adaptation performed on the end-user's machine as well (within the browser). Authoring of adaptive applications within the educational domain with the system proposed was tried out with first year students from the Design-Based Learning Hypermedia course at the Eindhoven University of Technology (TU/e) to gather feedback on the problems they faced with the platform. more...
- Published
- 2016
44. Adaptive web-based educational application for autistic students
- Author
-
García, AM, De Bra, P, Stash, N, Fletcher, GHL, Fabri, M, Pechenizkiy, M, García, AM, De Bra, P, Stash, N, Fletcher, GHL, Fabri, M, and Pechenizkiy, M
- Abstract
Adaptive web-based applications have proven successful in reducing navigation and comprehension problems in hypermedia documents. In this paper, we describe a toolkit that is offered as an adaptive Web-based application to help autistic students incorporate to high education. The toolkit has been developed using a popular CMS in which we have integrated a client-side adaptation library. The toolkit described here was tried out during workshops with autistic students at Leeds Becketts University to gather (mostly qualitative) feedback on the adaptation and privacy aspects of the Autism&Uni platform. That feedback was later used to improve the toolkit. more...
- Published
- 2016
45. ACLAC: An approach for adaptive closed-loop anesthesia control
- Author
-
Marrero, A., Mendez, J.A., Maslov, A., Pechenizkiy, M., Rodrigues, P.P., Gama, J., Information Systems WSK&I, and Data Mining
- Subjects
Identification (information) ,Adaptive control ,Control theory ,Remote patient monitoring ,business.industry ,Anesthesia ,Bispectral index ,Process (computing) ,Medicine ,business ,Change detection ,Data modeling - Abstract
In current practice, to control the anesthetic process, the anesthetist delivers drugs according to the surgery procedure and to the current patient characteristics and state. This is an open-loop procedure requiring an active participation of the medical expert. We propose an adaptive closed-loop controller for the regulation of hypnosis for patients undergoing general anesthesia. One of the main problems arising when designing such a controller is related to the intra- and inter-patient variability. We employ a simple regression model to make prediction of patient's response and to compute the adequate doses of propofol to keep the patient in the specified Bispectral Index target. To make our model adaptive, we continuously monitor the patient behavior and detect changes in patient response to update the identification model. Experimental evaluation on real patients data shows that we can effectively detect change points. Simulation of the adaptive closed-loop control with the change detection mechanism also suggests that the use of the adaptation mechanism improves the control. more...
- Published
- 2013
- Full Text
- View/download PDF
46. Proceedings of cmbs 2013 : 26th ieee international symposium on computer based medical systems, porto, 20-22 june university of porto
- Author
-
Rodrigues, P. Pereira, Pechenizkiy, M., Gama, J., Correia, R. Cruz, Liu, J., Traina, A., Lucas, P., and Soda, P.
- Subjects
Software Science - Abstract
Item does not contain fulltext 547 p.
- Published
- 2013
47. Proceedings of the 26th IEEE International Symposium on Computer-Based Medical Systems (CBMS'13, Porto, Portugal, June 20-22, 2013)
- Author
-
Pereira Rodrigues, P., Pechenizkiy, M., Gama, J., Cruz-Correia, R., Liu, Jiming, Traina, A.J.M., Lucas, P.J.F., Soda, P., and Data Mining
- Published
- 2013
48. Mobile sentiment analysis
- Author
-
Chambers, L., Tromp, E., Pechenizkiy, M., and Gaber, M.
- Subjects
Computing - Abstract
Mobile devices play a significant part in a user’s communication methods and much data that they read and write is received and sent via mobile phones, for instance SMS messages, e-mails, Twitter tweets and social media networking feeds. One of the main goals is to make people aware of how much negative and positive content they read and write via their mobile phones. Existing sentiment analysis applications perform sentiment analysis on downloaded data from mobile phones or use an application installed on another computer to perform the analysis. The sentiment analysis described in this paper is to be performed locally on the mobile phone enabling immediate and private analysis of personal messages and social media contents, allowing the users to be able to reason about their mood and stress level that may be affected by what they had been receiving. Experimental results showed the effectiveness of the proposed system on Android smartphones with varying computational capabilities. more...
- Published
- 2012
49. CurriM : Curriculum mining
- Author
-
Pechenizkiy, M., Trcka, N., De Bra, P.M.E., Toledo, Pedro A., Yacef, Kalina, Zaiane, Osmar R., Hershkovitz, Arnong, Yudelson, Michael, Electronic Systems, and Data Mining
- Subjects
ComputingMilieux_COMPUTERSANDEDUCATION - Abstract
Curriculum mining includes three main kinds of tasks: (i) actual curriculum model discovery, i.e. constructing complete and compact academic curriculum models that are able to reproduce the observed behavior of students, (ii) curriculum model conformance checking, i.e. checking whether the observed behavior of students match their expected behavior as defined by the previously discovered or pre-authored curriculum model, and (iii) curriculum model extension, i.e. projecting information extracted from the observed data onto the model, to make the tacit knowledge explicit, facilitate better understanding of the particular academic processes and enable decision making processes. We discuss student and education responsible perspectives on curriculum mining and present the achievements of the ongoing project aiming to develop curriculum mining software including process mining, data mining and visualization techniques. more...
- Published
- 2012
50. Stress analytics in education
- Author
-
Kocielnik, R.D., Pechenizkiy, M., Sidorova, N., Yacef, K., Zaïane, O.R., Hershkovitz, A., Yudelson, M., Stamper, J.C., Process Science, and Data Mining
- Subjects
education - Abstract
During the years of college and university education students are exposed to different kinds of stress, especially during the difficult studying periods like final exams weeks or project deadlines. Stress on a long run is dangerous and can contribute to illness through its physiological effects or maladaptive health behaviors. Many students admit, or are self-aware, that they become stressed under different circumstances and have some clues about their potential stressor. Still, even for such students, the monitoring and awareness of stress are not systematic and based on subjective data, i.e. someone's feelings. In our work we aim at providing means to students to become aware of the past, current and expected (objectively measured) stress and its correlation with their performance, to understand their stressors, to cope with and prevent stress - thus, to live healthier and happier lives and better organize their studies. more...
- Published
- 2012
Discovery Service for Jio Institute Digital Library
For full access to our library's resources, please sign in.