Abstract Text Classification is the process of grouping documents into categories based on their contents. Many Text Classification (TC) algorithms have been proposed for Arabic TC. The main idea in text Arabic classification is the accuracy of the results, which these results depends on the correctness of the text classification phases which start with the prepossessing phase and end with feature selection... Text Classification is the process of grouping documents into categories based on their contents. Many Text Classification (TC) algorithms have been proposed for Arabic TC. The main idea in text Arabic classification is the accuracy of the results, which these results depends on the correctness of the text classification phases which start with the prepossessing phase and end with feature selection and the choosing of the best classifier that can classify text in related groups. In this paper we provide a new system for text classification based on BPSO/REP-Tree hybrid. The first term refers to the “Binary Particle Swarm Optimization” that we use it for the feature selection process and the second term refers the classifier we used “Reduced Error Pruning Tree”. We will show the results of the experiments on a data-set collected from the BBC-Arabic website using the Weka tool which specific for data classification. Read More Read Less

Abstract This paper summarizes several lexical methods for more comprehensive affect recognition in text using an example of typed utterances. We introduce a set of algorithms that are capable of recognizing emotions of user’s statements in order to achieve more effective and smoother human-machine conversation. Aspects often neglected by existing systems working with Japanese language, e.g. compound sentences, double negation sentences, modifiers as adverbs and... This paper summarizes several lexical methods for more comprehensive affect recognition in text using an example of typed utterances. We introduce a set of algorithms that are capable of recognizing emotions of user’s statements in order to achieve more effective and smoother human-machine conversation. Aspects often neglected by existing systems working with Japanese language, e.g. compound sentences, double negation sentences, modifiers as adverbs and emoticons were combined and their higher effectiveness in recognizing affect in more complicated sentences was confirmed through evaluation experiments. The results are introduced together with separate analysis of emoticons’ influence on emotional load. We also discuss importance of predicting human emotions not only in the field of human-computer interaction but also its meaning for developing ethical chatbots. Read More Read Less

Abstract Providing the means for understanding natural language, or some groups of its realizations is one of the main goals of Artificial Intelligence (AI). For example, subfields of AI such as natural language processing ( NLP), or sentiment analysis (SA) focus on analyzing of speaker attitudes and emotions expressed in a sentence. A different kind of realization of natural language, which we focus on... Providing the means for understanding natural language, or some groups of its realizations is one of the main goals of Artificial Intelligence (AI). For example, subfields of AI such as natural language processing ( NLP), or sentiment analysis (SA) focus on analyzing of speaker attitudes and emotions expressed in a sentence. A different kind of realization of natural language, which we focus on in this research, is predicting trends of future events from available language data. It is a challenging task, since it focuses on obtaining information on future unfolding of events from the provided instances of language behavior. To make the task feasible we narrowed the scope of our research to detecting and extracting sentences which refer to future events as potentially the most useful in prediction of future unfolding of events. Future reference sentences can contain information on the current state of events, or specific expert knowledge, and thus can be useful in future prediction. We propose a method for automatic extraction of future reference sentences from news corpora using semantic and morphological information. We also perform an experiment in which the sentences extracted with the proposed method are applied to predict future unfolding of a set of specific events. Read More Read Less