Browsing by Subject "machine learning"

Now showing 1 - 4 of 4

Open Access
Batch learning of disjoint feature intervals
(1996) Akkuş, Aynur
This thesis presents several learning algorithms for multi-concept descriptions in the form of disjoint feature intervals, called Feature Interval Learning algorithms (FIL). These algorithms are batch supervised inductive learning algorithms, and use feature projections of the training instances for the representcition of the classification knowledge induced. These projections can be generalized into disjoint feature intervals. Therefore, the concept description learned is a set of disjoint intervals separately for each feature. The classification of an unseen instance is based on the weighted majority voting among the local predictions of features. In order to handle noisy instances, several extensions are developed by placing weights to intervals rather than features. Empirical evaluation of the FIL algorithms is presented and compared with some other similar classification algorithms. Although the FIL algorithms achieve comparable accuracies with other algorithms, their average running times are much more less than the others. This thesis also presents a new adaptation of the well-known /s-NN classification algorithm to the feature projections approach, called A:-NNFP for k-Nearest Neighbor on Feature Projections, based on a majority voting on individual classifications made by the projections of the training set on each feature and compares with the /:-NN algorithm on some real-world and cirtificial datasets.
Open Access
Benefit maximizing classification using feature intervals
(2002) İkizler, Nazlı
For a long time, classification algorithms have focused on minimizing the quantity of prediction errors by assuming that each possible error has identical consequences. However, in many real-world situations, this assumption is not convenient. For instance, in a medical diagnosis domain, misdiagnosing a sick patient as healthy is much more serious than its opposite. For this reason, there is a great need for new classification methods that can handle asymmetric cost and benefit constraints of classifications. In this thesis, we discuss cost-sensitive classification concepts and propose a new classification algorithm called Benefit Maximization with Feature Intervals (BMFI) that uses the feature projection based knowledge representation. In the framework of BMFI, we introduce five different voting methods that are shown to be effective over different domains. A number of generalization and pruning methodologies based on benefits of classification are implemented and experimented. Empirical evaluation of the methods has shown that BMFI exhibits promising performance results compared to recent wrapper cost-sensitive algorithms, despite the fact that classifier performance is highly dependent on the benefit constraints and class distributions in the domain. In order to evaluate costsensitive classification techniques, we describe a new metric, namely benefit accuracy which computes the relative accuracy of the total benefit obtained with respect to the maximum possible benefit achievable in the domain.
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
Classification with overlapping feature intervals
(1995) Koç, Hakime Ünsal
This thesis presents a new form of exemplar-based learning method, based on overlapping feature intervals. Classification with Overlapping Feature Intervals (COFI) is the particular implementation of this technique. In this incremental, inductive and supervised learning method, the basic unit of the representation is an interval. The COFI algorithm learns the projections of the intervals in each class dimension for each feature. An interval is initially a point on a class dimension, then it can be expanded through generalization. No specialization of intervals is done on class dimensions by this algorithm. Classification in the COFI algorithm is based on a majority voting among the local predictions that are made individually by each feature.