Đang chuẩn bị nút TẢI XUỐNG, xin hãy chờ
Tải xuống
This model identifies underlying semantic structure of a document collection from word to topic and topic to document quickly, effectively without human intervention. Initially, this model creates the topics based on Dirichlet Distribution using vector space model. | ISSN: 2249-5789 Adapala Spurthi, International Journal of Computer Science & Communication Networks,Vol 8(5),46-49 AN ENHANCED LEARNING TOPIC MODEL FOR EXPLOITING DOCUMENT RELATIONSHIPS USING CLASSIFICATION AND REGRESSION ADAPALA SPURTHI M. Tech, Department of CSE, Shri Vishnu Engineering College for Women (A), Vishnupur, Bhimavaram, West Godavari District, Andhra Pradesh Abstract- Massive influx of information technology produces high volumes of electronic documents and this repository has become potentially high dimensional, complex, and diversified in nature. The growing need of analyzing documents, discovering Latent data and identifying relationships among the documents in such large repositories, got the attention of present researchers towards the most powerful era of topic modeling. As the data in the repository growing dynamically and determined by higher variances, the conventional techniques modeled on Multinomial distributions, Dirichlet distributions in the topic models are limited in performing tasks like organizing, searching, and indexing. In handling such data and overcoming the limitations of previous techniques it is necessary to use the most popular method Latent Dirichlet Allocation (LDA) that can find the relationships among the data by representing as generative probabilistic distributions over latent topics. Towards this, the proposed work namely “AN Enhanced Learning Topic Model for Exploiting Document Relationships Using Classification and Regression” which is designed based on the principles of generative probabilistic topic models. This model identifies underlying semantic structure of a document collection from word to topic and topic to document quickly, effectively without human intervention. Initially, this model creates the topics based on Dirichlet Distribution using vector space model. Subsequently, the proposed model chooses the suitable annotators by distributing topics for each document to conduct fundamental validation in .
