Browsing by Subject "Learning systems"

Now showing 1 - 20 of 45

Open Access
Activity recognition invariant to sensor orientation with wearable motion sensors
(MDPI AG, 2017) Yurtman, A.; Barshan, B.
Most activity recognition studies that employ wearable sensors assume that the sensors are attached at pre-determined positions and orientations that do not change over time. Since this is not the case in practice, it is of interest to develop wearable systems that operate invariantly to sensor position and orientation. We focus on invariance to sensor orientation and develop two alternative transformations to remove the effect of absolute sensor orientation from the raw sensor data. We test the proposed methodology in activity recognition with four state-of-the-art classifiers using five publicly available datasets containing various types of human activities acquired by different sensor configurations. While the ordinary activity recognition system cannot handle incorrectly oriented sensors, the proposed transformations allow the sensors to be worn at any orientation at a given position on the body, and achieve nearly the same activity recognition performance as the ordinary system for which the sensor units are not rotatable. The proposed techniques can be applied to existing wearable systems without much effort, by simply transforming the time-domain sensor data at the pre-processing stage. © 2017 by the authors. Licensee MDPI, Basel, Switzerland.
Open Access
Adaptive ensemble learning with confidence bounds for personalized diagnosis
(AAAI Press, 2016) Tekin, Cem; Yoon, J.; Van Der Schaar, M.
With the advances in the field of medical informatics, automated clinical decision support systems are becoming the de facto standard in personalized diagnosis. In order to establish high accuracy and confidence in personalized diagnosis, massive amounts of distributed, heterogeneous, correlated and high-dimensional patient data from different sources such as wearable sensors, mobile applications, Electronic Health Record (EHR) databases etc. need to be processed. This requires learning both locally and globally due to privacy constraints and/or distributed nature of the multimodal medical data. In the last decade, a large number of meta-learning techniques have been proposed in which local learners make online predictions based on their locally-collected data instances, and feed these predictions to an ensemble learner, which fuses them and issues a global prediction. However, most of these works do not provide performance guarantees or, when they do, these guarantees are asymptotic. None of these existing works provide confidence estimates about the issued predictions or rate of learning guarantees for the ensemble learner. In this paper, we provide a systematic ensemble learning method called Hedged Bandits, which comes with both long run (asymptotic) and short run (rate of learning) performance guarantees. Moreover, we show that our proposed method outperforms all existing ensemble learning techniques, even in the presence of concept drift.
Open Access
Adaptive hierarchical space partitioning for online classification
(IEEE, 2016) Kılıç, O. Fatih; Vanlı, N. D.; Özkan, H.; Delibalta, İ.; Kozat, Süleyman Serdar
We propose an online algorithm for supervised learning with strong performance guarantees under the empirical zero-one loss. The proposed method adaptively partitions the feature space in a hierarchical manner and generates a powerful finite combination of basic models. This provides algorithm to obtain a strong classification method which enables it to create a linear piecewise classifier model that can work well under highly non-linear complex data. The introduced algorithm also have scalable computational complexity that scales linearly with dimension of the feature space, depth of the partitioning and number of processed data. Through experiments we show that the introduced algorithm outperforms the state-of-the-art ensemble techniques over various well-known machine learning data sets.
Open Access
Application of the RIMARC algorithm to a large data set of action potentials and clinical parameters for risk prediction of atrial fibrillation
(Springer, 2015) Ravens, U.; Katircioglu-Öztürk, D.; Wettwer, E.; Christ, T.; Dobrev, D.; Voigt, N.; Poulet, C.; Loose, S.; Simon, J.; Stein, A.; Matschke, K.; Knaut, M.; Oto, E.; Oto, A.; Güvenir, H. A.
Ex vivo recorded action potentials (APs) in human right atrial tissue from patients in sinus rhythm (SR) or atrial fibrillation (AF) display a characteristic spike-and-dome or triangular shape, respectively, but variability is huge within each rhythm group. The aim of our study was to apply the machine-learning algorithm ranking instances by maximizing the area under the ROC curve (RIMARC) to a large data set of 480 APs combined with retrospectively collected general clinical parameters and to test whether the rules learned by the RIMARC algorithm can be used for accurately classifying the preoperative rhythm status. APs were included from 221 SR and 158 AF patients. During a learning phase, the RIMARC algorithm established a ranking order of 62 features by predictive value for SR or AF. The model was then challenged with an additional test set of features from 28 patients in whom rhythm status was blinded. The accuracy of the risk prediction for AF by the model was very good (0.93) when all features were used. Without the seven AP features, accuracy still reached 0.71. In conclusion, we have shown that training the machine-learning algorithm RIMARC with an experimental and clinical data set allows predicting a classification in a test data set with high accuracy. In a clinical setting, this approach may prove useful for finding hypothesis-generating associations between different parameters.
Open Access
Authorship attribution: performance of various features and classification methods
(IEEE, 2007-11) Bozkurt, İlker Nadi; Bağlıoğlu, Özgür; Uyar, Erkan
Authorship attribution is the process of determining the writer of a document. In literature, there are lots of classification techniques conducted in this process. In this paper we explore information retrieval methods such as tf-idf structure with support vector machines, parametric and nonparametric methods with supervised and unsupervised (clustering) classification techniques in authorship attribution. We performed various experiments with articles gathered from Turkish newspaper Milliyet. We performed experiments on different features extracted from these texts with different classifiers, and combined these results to improve our success rates. We identified which classifiers give satisfactory results on which feature sets. According to experiments, the success rates dramatically changes with different combinations, however the best among them are support vector classifier with bag of words, and Gaussian with function words. ©2007 IEEE.
Open Access
Automatic categorization of ottoman literary texts by poet and time period
(Springer, London, 2012) Can, Ethem F.; Can, Fazlı; Duygulu, Pınar; Kalpaklı, Mehmet
Millions of manuscripts and printed texts are available in the Ottoman language. The automatic categorization of Ottoman texts would make these documents much more accessible in various applications ranging from historical investigations to literary analyses. In this work, we use transcribed version of Ottoman literary texts in the Latin alphabet and show that it is possible to develop effective Automatic Text Categorization techniques that can be applied to the Ottoman language. For this purpose, we use two fundamentally different machine learning methods: Naïve Bayes and Support Vector Machines, and employ four style markers: most frequent words, token lengths, two-word collocations, and type lengths. In the experiments, we use the collected works (divans) of ten different poets: two poets from five different hundred-year periods ranging from the 15th to 19th century. The experimental results show that it is possible to obtain highly accurate classifications in terms of poet and time period. By using statistical analysis we are able to recommend which style marker and machine learning method are to be used in future studies. © 2012 Springer-Verlag London Limited.
Open Access
Big-data streaming applications scheduling based on staged multi-armed bandits
(Institute of Electrical and Electronics Engineers, 2016) Kanoun, K.; Tekin, C.; Atienza, D.; Van Der Schaar, M.
Several techniques have been recently proposed to adapt Big-Data streaming applications to existing many core platforms. Among these techniques, online reinforcement learning methods have been proposed that learn how to adapt at run-time the throughput and resources allocated to the various streaming tasks depending on dynamically changing data stream characteristics and the desired applications performance (e.g., accuracy). However, most of state-of-the-art techniques consider only one single stream input in its application model input and assume that the system knows the amount of resources to allocate to each task to achieve a desired performance. To address these limitations, in this paper we propose a new systematic and efficient methodology and associated algorithms for online learning and energy-efficient scheduling of Big-Data streaming applications with multiple streams on many core systems with resource constraints. We formalize the problem of multi-stream scheduling as a staged decision problem in which the performance obtained for various resource allocations is unknown. The proposed scheduling methodology uses a novel class of online adaptive learning techniques which we refer to as staged multi-armed bandits (S-MAB). Our scheduler is able to learn online which processing method to assign to each stream and how to allocate its resources over time in order to maximize the performance on the fly, at run-time, without having access to any offline information. The proposed scheduler, applied on a face detection streaming application and without using any offline information, is able to achieve similar performance compared to an optimal semi-online solution that has full knowledge of the input stream where the differences in throughput, observed quality, resource usage and energy efficiency are less than 1, 0.3, 0.2 and 4 percent respectively.
Open Access
Boosted LMS-based piecewise linear adaptive filters
(IEEE, 2016) Kari, Dariush; Marivani, Iman; Delibalta, İ.; Kozat, Süleyman Serdar
We introduce the boosting notion extensively used in different machine learning applications to adaptive signal processing literature and implement several different adaptive filtering algorithms. In this framework, we have several adaptive constituent filters that run in parallel. For each newly received input vector and observation pair, each filter adapts itself based on the performance of the other adaptive filters in the mixture on this current data pair. These relative updates provide the boosting effect such that the filters in the mixture learn a different attribute of the data providing diversity. The outputs of these constituent filters are then combined using adaptive mixture approaches. We provide the computational complexity bounds for the boosted adaptive filters. The introduced methods demonstrate improvement in the performances of conventional adaptive filtering algorithms due to the boosting effect.
Open Access
Çağrı merkezi metin madenciliği yaklaşımı
(IEEE, 2017-05) Yiğit, İ. O.; Ateş, A. F.; Güvercin, Mehmet; Ferhatosmanoğlu, Hakan; Gedik, Buğra
Günümüzde çağrı merkezlerindeki görüşme kayıtlarının sesten metne dönüştürülebilmesi görüşme kaydı metinleri üzerinde metin madenciliği yöntemlerinin uygulanmasını mümkün kılmaktadır. Bu çalışma kapsamında görüşme kaydı metinleri kullanarak görüşmenin içeriğinin duygu yönünden (olumlu/olumsuz) değerlendirilmesi, müşteri memnuniyetinin ve müşteri temsilcisi performansının ölçülmesi amaçlanmaktadır. Yapılan çalışmada görüşme kaydı metinlerinden metin madenciliği yöntemleri ile yeni özellikler çıkarılmıştır. Metinlerden elde edilen özelliklerden yararlanılarak sınıflandırma ve regresyon yöntemleriyle görüşme kayıtlarının içeriklerinin değerlendirilmesini sağlayacak tahmin modelleri oluşturulmuştur. Bu çalışma sonucunda ortaya çıkarılan tahmin modellerinin Türk Telekom bünyesindeki çağrı merkezlerinde kullanılması hedeflenmektedir.
Open Access
Cell-graph mining for breast tissue modeling and classification
(IEEE, 2007-08) Bilgin, C.; Demir, Çiğdem; Nagi, C.; Yener, B.
We consider the problem of automated cancer diagnosis in the context of breast tissues. We present graph theoretical techniques that identify and compute quantitative metrics for tissue characterization and classification. We segment digital images of histopatological tissue samples using k-means algorithm. For each segmented image we generate different cell-graphs using positional coordinates of cells and surrounding matrix components. These cell-graphs have 500-2000 cells(nodes) with 1000-10000 links depending on the tissue and the type of cell-graph being used. We calculate a set of global metrics from cell-graphs and use them as the feature set for learning. We compare our technique, hierarchical cell graphs, with other techniques based on intensity values of images, Delaunay triangulation of the cells, the previous technique we proposed for brain tissue images and with the hybrid approach that we introduce in this paper. Among the compared techniques, hierarchical-graph approach gives 81.8% accuracy whereas we obtain 61.0%, 54.1% and 75.9% accuracy with intensity-based features, Delaunay triangulation and our previous technique, respectively. © 2007 IEEE.
Open Access
Classification by voting feature intervals
(Springer, 1997-04) Demiröz, Gülşen; Güvenir, H. Altay
A new classification algorithm called VFI (for Voting Feature Intervals) is proposed. A concept is represented by a set of feature intervals on each feature dimension separately. Each feature participates in the classification by distributing real-valued votes among classes. The class receiving the highest vote is declared to be the predicted class. VFI is compared with the Naive Bayesian Classifier, which also considers each feature separately. Experiments on real-world datasets show that VFI achieves comparably and even better than NBC in terms of classification accuracy. Moreover, VFI is faster than NBC on all datasets. © Springer-Verlag Berlin Heidelberg 1997.
Open Access
Classification of regional ionospheric disturbance based on machine learning techniques
(European Space Agency, 2016) Terzi, Merve Begüm; Arıkan, Orhan; Karatay, S.; Arıkan, F.; Gulyaeva, T.
In this study, Total Electron Content (TEC) estimated from GPS receivers is used to model the regional and local variability that differs from global activity along with solar and geomagnetic indices. For the automated classification of regional disturbances, a classification technique based on a robust machine learning technique that have found wide spread use, Support Vector Machine (SVM) is proposed. Performance of developed classification technique is demonstrated for midlatitude ionosphere over Anatolia using TEC estimates generated from GPS data provided by Turkish National Permanent GPS Network (TNPGN-Active) for solar maximum year of 2011. As a result of implementing developed classification technique to Global Ionospheric Map (GIM) TEC data, which is provided by the NASA Jet Propulsion Laboratory (JPL), it is shown that SVM can be a suitable learning method to detect anomalies in TEC variations.
Open Access
Computer vision based text and equation editor for LATEX
(IEEE, 2004-06) Öksüz, Özcan; Güdükbay, Uğur; Çetin, Enis
In this paper, we present a computer vision based text and equation editor for LATEX. The user writes text and equations on paper and a camera attached to a computer records actions of the user. In particular, positions of the pen-tip in consecutive image frames are detected. Next, directional and positional information about characters are calculated using these positions. Then, this information is used for on-line character classification. After characters and symbols are found, corresponding LATEX code is generated.
Open Access
Data imputation through the identification of local anomalies
(Institute of Electrical and Electronics Engineers Inc., 2015) Ozkan, H.; Pelvan, O. S.; Kozat, S. S.
We introduce a comprehensive and statistical framework in a model free setting for a complete treatment of localized data corruptions due to severe noise sources, e.g., an occluder in the case of a visual recording. Within this framework, we propose: 1) a novel algorithm to efficiently separate, i.e., detect and localize, possible corruptions from a given suspicious data instance and 2) a maximum a posteriori estimator to impute the corrupted data. As a generalization to Euclidean distance, we also propose a novel distance measure, which is based on the ranked deviations among the data attributes and empirically shown to be superior in separating the corruptions. Our algorithm first splits the suspicious instance into parts through a binary partitioning tree in the space of data attributes and iteratively tests those parts to detect local anomalies using the nominal statistics extracted from an uncorrupted (clean) reference data set. Once each part is labeled as anomalous versus normal, the corresponding binary patterns over this tree that characterize corruptions are identified and the affected attributes are imputed. Under a certain conditional independency structure assumed for the binary patterns, we analytically show that the false alarm rate of the introduced algorithm in detecting the corruptions is independent of the data and can be directly set without any parameter tuning. The proposed framework is tested over several well-known machine learning data sets with synthetically generated corruptions and experimentally shown to produce remarkable improvements in terms of classification purposes with strong corruption separation capabilities. Our experiments also indicate that the proposed algorithms outperform the typical approaches and are robust to varying training phase conditions. © 2015 IEEE.
Open Access
Distance-based classification methods
(Taylor & Francis, 1999) Ekin, O.; Hammer, P. L.; Kogan, A.; Winter, P.
Given a set of points in a Euclidean space, and a partitioning of this 'training set' into two or more subsets ('classes'), we consider the problem of identifying a 'reasonable' assignment of another point in the Euclidean space ('query point') to one of these classes. The various classifications proposed in this paper are determined by the distances between the query point and the points in the training set. We report results of extensive computational experiments comparing the new methods with two well-known distance-based classification methods (k-nearest neighbors and Parzen windows) on data sets commonly used in the literature. The results show that the performance of both new and old distance-based methods is on par with and often better than that of the other best classification methods known. Moreover, the new classification procedures proposed in this paper are: (i) easy to implement, (ii) extremely fast, and (iii) very robust (i.e. their performance is insignificantly affected by the choice of parameter values).
Open Access
The effect of uncertainty on learning in game-like environments
(Pergamon Press, 2013) Ozcelik, E.; Cagiltay, N. E.; Ozcelik, N. S.
Considering the role of games for educational purposes, there has an increase in interest among educators in applying strategies used in popular games to create more engaging learning environments. Learning is more fun and appealing in digital educational games and, as a result, it may become more effective. However, few research studies have been conducted to establish principles based on empirical research for designing engaging and entertaining games so as to improve learning. One of the essential characteristics of games that has been unexplored in the literature is the concept of uncertainty. This study examines the effect of uncertainty on learning outcomes. In order to better understand this effect on learning, a game-like learning tool was developed to teach a database concept in higher education programs of software engineering. The tool is designed in two versions: one including uncertainty and the other including no uncertainty. The experimental results of this study reveal that uncertainty enhances learning. Uncertainty is found to be positively associated with motivation. As motivation increases, participants tend to spend more time on answering the questions and to have higher accuracy in these questions. © 2013 Elsevier Ltd. All rights reserved.
Open Access
An energy efficient additive neural network
(IEEE, 2017) Afrasiyabi, A.; Nasir, B.; Yıldız, O.; Yarman-Vural, F. T.; Çetin, A. Enis
In this paper, we propose a new energy efficient neural network with the universal approximation property over space of Lebesgue integrable functions. This network, called additive neural network, is very suitable for mobile computing. The neural structure is based on a novel vector product definition, called ef-operator, that permits a multiplier-free implementation. In ef-operation, the 'product' of two real numbers is defined as the sum of their absolute values, with the sign determined by the sign of the product of the numbers. This 'product' is used to construct a vector product in n-dimensional Euclidean space. The vector product induces the lasso norm. The proposed additive neural network successfully solves the XOR problem. The experiments on MNIST dataset show that the classification performances of the proposed additive neural networks are very similar to the corresponding multi-layer perceptron.
Open Access
Ensemble of multiple instance classifiers for image re-ranking
(Elsevier Ltd, 2014) Sener F.; Ikizler-Cinbis, N.
Text-based image retrieval may perform poorly due to the irrelevant and/or incomplete text surrounding the images in the web pages. In such situations, visual content of the images can be leveraged to improve the image ranking performance. In this paper, we look into this problem of image re-ranking and propose a system that automatically constructs multiple candidate "multi-instance bags (MI-bags)", which are likely to contain relevant images. These automatically constructed bags are then utilized by ensembles of Multiple Instance Learning (MIL) classifiers and the images are re-ranked according to the final classification responses. Our method is unsupervised in the sense that, the only input to the system is the text query itself, without any user feedback or annotation. The experimental results demonstrate that constructing multiple instance bags based on the retrieval order and utilizing ensembles of MIL classifiers greatly enhance the retrieval performance, achieving on par or better results compared to the state-of-the-art. © 2014 Elsevier B.V.
Open Access
Extraction of sparse spatial filters using Oscillating Search
(IEEE, 2012) Onaran, İbrahim; İnce, N. Fırat; Abosch, A.; Çetin, A. Enis
Common Spatial Pattern algorithm (CSP) is widely used in Brain Machine Interface (BMI) technology to extract features from dense electrode recordings by using their weighted linear combination. However, the CSP algorithm, is sensitive to variations in channel placement and can easily overfit to the data when the number of training trials is insufficient. Construction of sparse spatial projections where a small subset of channels is used in feature extraction, can increase the stability and generalization capability of the CSP method. The existing 0 norm based sub-optimal greedy channel reduction methods are either too complex such as Backward Elimination (BE) which provided best classification accuracies or have lower accuracy rates such as Recursive Weight Elimination (RWE) and Forward Selection (FS) with reduced complexity. In this paper, we apply the Oscillating Search (OS) method which fuses all these greedy search techniques to sparsify the CSP filters. We applied this new technique on EEG dataset IVa of BCI competition III. Our results indicate that the OS method provides the lowest classification error rates with low cardinality levels where the complexity of the OS is around 20 times lower than the BE. © 2012 IEEE.
Open Access
Generalizing predicates with string arguments
(Springer New York LLC, 2006-06) Cicekli, I.; Cicekli, N. K.
The least general generalization (LGG) of strings may cause an over-generalization in the generalization process of the clauses of predicates with string arguments. We propose a specific generalization (SG) for strings to reduce over-generalization. SGs of strings are used in the generalization of a set of strings representing the arguments of a set of positive examples of a predicate with string arguments. In order to create a SG of two strings, first, a unique match sequence between these strings is found. A unique match sequence of two strings consists of similarities and differences to represent similar parts and differing parts between those strings. The differences in the unique match sequence are replaced to create a SG of those strings. In the generalization process, a coverage algorithm based on SGs of strings or learning heuristics based on match sequences are used. © Springer Science + Business Media, LLC 2006.