Abstract In this paper, a deep learning RNN model is used to classify Tibetan texts. The core idea first preprocess the news corpus, and then use syllables construct syllable table based on lexical grammatical structure of Tibetan, embed in sentence, represent each as fixed Numerical vector. Secondly, cyclic neural network constructed. First, text different lengths filled or truncated into sequ...

Micro blogging tools provide a real time service for the public to express opinions, broadcast news and information offer an opportunity comment respond such output. Word usage in social media is continually evolving. bloggers may use different sets of words describe specific event they new (i.e. neither exist training dataset nor informal or formal dictionaries) contexts. Dynamically capturing...

In this project report, I will evaluate the several text clustering approaches and how they can be used for the purpose of text classification. The particular task is topic classification of 20 Newsgroup dataset and sentiment classification restaurant reviews dataset. Future direction for improving the results will also be discussed.

We describe a multiclass text classification system for formatted text messages contained in the Rich Text Format fields of a structured database of military documents. This system uses a Part-Of–Speech tagger and a RuleBased Classifier to classify 80 different types of formatted messages.

In this paper we present Document analysis and classification system to segment and classify contents of Arabic document images. This system includes preprocessing, document segmentation, feature extraction and document classification. A document image is enhanced in the preprocessing by removing noise, binarization, and detecting and correcting image skew. In document segmentation, an algorith...

Distance to Transition Point (DTP) has shown good performance in term selection for Text Categorization task. Previous experiment report that DTP behaves well as DF and CHI term selection techniques. In this paper we present the results of using DTP computed in a global and local fashion; considering the whole of categories of training set. The results confirm that performance of DTP globally c...