In this study, support vector regression (SVR) analysis is used as a machine learning technique in order to predict the stock market price as well as to predict stock market trend. Moreover, different types of windowing operators are used as data preprocess or input selection technique for SVR models. This is a new approach which uses different types of windowing functions as data preprocess for predicting time series data. Support vector regression is a useful and powerful machine learning technique to recognize pattern of time series dataset. It can produce good prediction result if the value of important parameters can be determined properly. Different kinds of Windowing operators are used in this experiment in order to feed more reliable inputs into regression models. This study is done on a well known company of Dhaka stock exchange (DSE), named ACI group of company Limited. Four year's historical time series dataset are collected from the DSE from 2009 to 2012, as daily basis for experimentations. Finally, predicted results from WinSVR models are compared with actual price values of DSE to evaluate the model prediction performance.
Question classification is a crucial task for answer selection. Question classification could help define the structure of question sentences generated by features extraction from a sentence, such as who, when, where, and how. In this paper, we proposed a methodology to improve question classification from texts by using feature selection and word embedding techniques. We conducted several experiments to evaluate the performance of the proposed methodology using two different datasets (TREC-6 dataset and Thai sentence dataset) with term frequency and combined term frequency-inverse document frequency including Unigram, Unigram+Bigram, and Unigram + Trigram as features. Machine learning models based on traditional and deep learning classifiers were used. The traditional classification models were Multinomial Naïve Bayes, Logistic Regression, and Support Vector Machine. The deep learning techniques were Bidirectional Long Short-Term Memory (BiLSTM), Convolutional Neural Networks (CNN), and Hybrid model, which combined CNN and BiLSTM model. The experiment results showed that our methodology based on Part-of-Speech (POS) tagging was the best to improve question classification accuracy. The classifying question categories achieved with average micro -score of 0.98 when applied SVM model on adding all POS tags in the TREC-6 dataset. The highest average micro -score achieved 0.8 when applied GloVe by using CNN model on adding focusing tags in the Thai sentences dataset.