A novel field learning algorithm for dual imbalance text classification

Fish-net algorithm is a novel field learning algorithm which derives classification rules by looking at the range of values of each attribute instead of the individual point values. In this paper, we present a Feature Selection Fish-net learning algorithm to solve the Dual Imbalance problem on text classification. Dual imbalance includes the instance imbalance and feature imbalance. The instance imbalance is caused by the unevenly distributed classes and feature imbalance is due to the different document length. The proposed approach consists of two phases: (1) select a feature subset which consists of the features that are more supportive to difficult minority class; (2) construct classification rules based on the original Fish-net algorithm. Our experimental results on Reuters21578 show that the proposed approach achieves better balanced accuracy rate on both majority and minority class than Naive Bayes MultiNomial and SVM.

Evolutionary structure learning algorithm for Bayesian network and penalized mutual information metric

This paper formulates the problem of learning Bayesian network structures from data as determining the structure that best approximates the probability distribution indicated by the data. A new metric, Penalized Mutual Information metric, is proposed, and a evolutionary algorithm is designed to search for the best structure among alternatives. The experimental results show that this approach is reliable and promising.

A neural networks learning algorithm for minor component analysis and its convergence analysis

The eigenvector associated with the smallest eigenvalue of the autocorrelation matrix of input signals is called minor component. Minor component analysis (MCA) is a statistical approach for extracting minor component from input signals and has been applied in many fields of signal processing and data analysis. In this letter, we propose a neural networks learning algorithm for estimating adaptively minor component from input signals. Dynamics of the proposed algorithm are analyzed via a deterministic discrete time (DDT) method. Some sufficient conditions are obtained to guarantee convergence of the proposed algorithm.

On the discrete time dynamics of a self-stabilizing MCA learning algorithm

The stability of minor component analysis (MCA) learning algorithms is an important problem in many signal processing applications. In this paper, we propose an effective MCA learning algorithm that can offer better stability. The dynamics of the proposed algorithm are analyzed via a corresponding deterministic discrete time (DDT) system. It is proven that if the learning rate satisfies some mild conditions, almost all trajectories of the DDT system starting from points in an invariant set are bounded, and will converge to the minor component of the autocorrelation matrix of the input data. Simulation results will be furnished to illustrate the theoretical results achieved.

A stable MCA learning algorithm

Minor component analysis (MCA) is an important statistical tool for signal processing and data analysis. Neural networks can be used to extract online minor component from input data. Compared with traditional algebraic  approaches, a neural network method has a lower computational complexity. Stability of neural networks learning algorithms is crucial to practical applications. In this paper, we propose a stable MCA neural networks learning algorithm, which has a more satisfactory numerical stability than some existing MCA algorithms. Dynamical behaviors of the proposed algorithm are analyzed via deterministic discrete time (DDT) method and the conditions are obtained to guarantee convergence. Simulations are carried out to illustrate the theoretical results achieved.

A unified learning algorithm to extract principal and minor components

Recently, many unified learning algorithms have been developed to solve the task of principal component analysis (PCA) and minor component analysis (MCA). These unified algorithms can be used to extract principal component and if altered simply by the sign, it can also serve as a minor component extractor. This is of practical significance in the implementations of algorithms. Convergence of the existing unified algorithms is guaranteed only under the condition that the learning rates of algorithms approach zero, which is impractical in many practical applications. In this paper, we propose a unified PCA & MCA algorithm with a constant learning rate, and derive the sufficient conditions to guarantee convergence via analyzing the discrete-time dynamics of the proposed algorithm. The achieved theoretical results lay a solid foundation for the applications of our proposed algorithm.

On the incremental learning and recognition of the pattern of movement of multiple labelled objects in dynamic scenes

In this paper we discuss combining incremental learning and incremental recognition to classify patterns consisting of multiple objects, each represented by multiple spatio-temporal features. Importantly the technique allows for ambiguity in terms of the positions of the start and finish of the pattern. This involves a progressive classification which considers the data at each time instance in the query and thus provides a probable answer before all the query information becomes available. We present two methods that combine incremental learning and incremental recognition: a time instance method and an overall best match method.

An application of machine learning techniques for the classification of glaucomatous progression

This paper presents an application of machine learning to the problem of classifying patients with glaucoma into one of two classes:stable and progressive glaucoma. The novelty of the work is the use of new features for the data analysis combined with machine learning techniques to classify the medical data. The paper describes the new features and the results of using decision trees to separate stable and progressive cases. Furthermore, we show the results of using an incremental learning algorithm for tracking stable and progressive cases over time. In both cases we used a dataset of progressive and stable glaucoma patients obtained from a glaucoma clinic.

Classifying and learning cricket shots using camera motion

This paper presents a method to classify and learn cricket shots. The procedure begins by extracting the camera motion parameters from the shots. Then the camera parameter values are converted to symbolic form and combined to generate a symbolic description that defines the trajectory of the cricket bell. The description generated is used to classify the cricket shot and to dynamically expand or update the system's knowledge of shots. The first novel aspect of this approach is that by using the camera motion parameters, a complex and difficult process of low level image segmenting of either the batsman or the cricket ball from video images is avoided. Also the method does not require high resolution images. Another novel aspect of this work is the use of a new incremental learning algorithm that enables the system to improve and update its knowledge base. Unlike previously developed algorithms which store training instances and have simple method to prime their concept hierarchies, the incremental learning algorithm used in this work generates compact concept hierarchies and uses evidence based forgetting. The results show that the system performs well in the task of classifying four types of cricket shots.

Incremental learning of temporally-coherent gaussian mixture models

In this paper we address the problem of learning Gaussian Mixture Models (GMMs) incrementally. Unlike previous approaches which universally assume that new data comes in blocks representable by GMMs which are then merged with the current model estimate, our method works for the case when novel data points arrive oneby- one, while requiring little additional memory. We keep only two GMMs in the memory and no historical data. The current fit is updated with the assumption that the number of components is fixed, which is increased (or reduced) when enough evidence for a new component is seen. This is deduced from the change from the oldest fit of the same complexity, termed the Historical GMM, the concept of which is central to our method. The performance of the proposed method is demonstrated qualitatively and quantitatively on several synthetic data sets and video sequences of faces acquired in realistic imaging conditions

Incremental learning of temporally-coherent Gaussian mixture models

In this paper we address the problem of learning Gaussian Mixture Models (GMMs) incrementally. Unlike previous approaches which universally assume that new data comes in blocks representable by GMMs which are then merged with the current model estimate, our method works for the case when novel data points arrive one- by-one, while requiring little additional memory. We keep only two GMMs in the memory and no historical data. The current fit is updated with the assumption that the number of components is fixed which is increased (or reduced) when enough evidence for a new component is seen. This is deducedfrom the change from the oldest fit of the same complexity, termed the Historical GMM, the concept of which is central to our method. The performance of the proposed method is demonstrated qualitatively and quantitatively on several synthetic data sets and video sequences of faces acquired in realistic imaging conditions.

Using multiple windows to track concept drift

In this paper we present a multiple window incremental learning algorithm that distinguishes between virtual concept drift and real concept drift. The algorithm is unsupervised and uses a novel approach to tracking concept drift that involves the use of competing windows to interpret the data. Unlike previous methods which use a single window to determine the drift in the data, our algorithm uses three windows of different sizes to estimate the change in the data. The advantage of this approach is that it allows the system to progressively adapt and predict the change thus enabling it to deal more effectively with different types of drift. We give a detailed description of the algorithm and present the results obtained from its application to two real world problems: background image processing and sound recognition. We also compare its performance with FLORA, an existing concept drift tracking algorithm.