This paper describes the Bangla Document Categorization using Stochastic Gradient Descent (SGD) classifier. Here, document categorization is the task in which text documents are classified into one or more of predefined categories based on their contents. The proposed system can be divided into three steps: 1. feature extraction incorporating term frequency (TF) and inverse document frequency (IDF), 2. classifier design using the Stochastic Gradient Descent (SGD) algorithm by learning the distinct features, and 3. performance measure using F1-score. In the experiments on BDNews24 documents, it is observed that our proposed method provides higher accuracy in comparison with the methods based on Support Vector Machine (SVM) and Naive Bayesian (NB) classifier.

No abstract

No abstract

Lossless text data compression is an important field as it significantly reduces storage requirement and communication cost. In this work, the focus is directed mainly to different file compression coding techniques and comparisons between them. Some memory efficient encoding schemes are analyzed and implemented in this work. They are: Shannon Fano Coding, Huffman Coding, Repeated Huffman Coding and Run-Length coding. A new algorithm "Modified Run-Length Coding" is also proposed and compared with the other algorithms. These analyses show how these coding techniques work, how much compression is possible for these coding techniques, the amount of memory needed for each technique, comparison between these techniques to find out which technique is better in what conditions. It is observed from the experiments that the repeated Huffman Coding shows higher compression ratio. Besides, the proposed Modified run length coding shows a higher performance than the conventional one.

This chapter presents Bangla (widely known as Bengali) Automatic Speech Recognition (ASR) techniques by evaluating the different speech features, such as Mel Frequency Cepstral Coefficients (MFCCs), Local Features (LFs), phoneme probabilities extracted by time delay artificial neural networks of different architectures. Moreover, canonicalization of speech features is also performed for Gender-Independent (GI) ASR. In the canonicalization process, the authors have designed three classifiers by male, female, and GI speakers, and extracted the output probabilities from these classifiers for measuring the maximum. The maximization of output probabilities for each speech file provides higher correctness and accuracies for GI speech recognition. Besides, dynamic parameters (velocity and acceleration coefficients) are also used in the experiments for obtaining higher accuracy in phoneme recognition. From the experiments, it is also shown that dynamic parameters with hybrid features also increase the phoneme recognition performance in a certain extent. These parameters not only increase the accuracy of the ASR system, but also reduce the computation complexity of Hidden Markov Model (HMM)-based classifiers with fewer mixture components.