مهسا اذرشب

Nowadays, with the advancement of technology, a large amount of unstructured data is being generated. Textual data account for a large proportion of unstructured human-generated data. Therefore, because the number of digital documents is increasing rapidly, an automatic organization (without human intervention and interaction) is vital and valuable for these documents. Text clustering, as a process of dividing a collection of documents based on their similarity to separate subsets, is one of the most useful and widely used methods of organizing these documents. One of the factors affecting the quality and accuracy of clustering is the size of these documents. In fact, text documents contain scattered and uninformative features that affect the effectiveness of the text clustering method. The feature selection technique can be used to remove uninformative features and improve text clustering performance. The problem of feature selection and text clustering can be considered as an optimization problem and meta-heuristic algorithms can be used to solve it. In this research, a hybrid algorithm based on teaching-learing-based optimization and grey wolf optimization is proposed in order to enhance the feature selection and text clustering performance. To eva‎luate the proposed algorithm, several optimization algorithms, which are recently used to solve the same issue are simulated. The results indicate that the proposed algorithm makes a significant improvement by considering the common clustering eva‎luation measures.

خوشه¬بندي متن , انتخاب ويژگي , الگوريتم بهينه سازي مبتني بر آموزش و يادگيري , الگوريتم گرگ خاكستري , الگوريتم ژنتيك

Text clustering , feature selection , teaching-learning-based optimization , grey wolf optimization , genetic algorithm

mahsa azarshab

mohammad fathian