Dissertation > Excellent graduate degree dissertation topics show

Fine-Grained Sentiment Analysis Oriented in Product Domain

Author: WangShanYu
Tutor: ZhengDeQuan
School: Harbin Institute of Technology
Course: Computer Science and Technology
Keywords: Text Sentiment Analysis Emotional dictionary Product attribute extraction Field transplant CRFs
CLC: TP391.1
Type: Master's thesis
Year: 2011
Downloads: 143
Quote: 0
Read: Download Dissertation

Abstract


With the development of network information, network emerged a large number of user comments information, and how these texts collate and analyze emotional information extracted user concerns aroused widespread interest in research workers. Text sentiment analysis as a natural language processing an application direction gradually become a new hot spot. Text sentiment analysis, also known as opinion mining, refers to having emotional subjectivity text analysis, process, summarize and reasoning process. Text sentiment analysis involves a number of challenging research tasks, according to the research content of different text sentiment analysis can be classified into emotional, affective and emotional information extraction such as information retrieval and induction. Extraction of emotional information which is valuable text extraction emotion emotional information, emotional information extraction has practical value. Sentiment analysis around the text, the text of the work consisted primarily of the following elements: (a) the construction of technical texts emotional resources. In order to solve the current problem of lack of emotional resources, text drawing existing research results and the use of three methods for semantic lexicon expansion, expanding the scale of emotional resources to enhance the credibility of emotional resources, and expanded the emotional words Dictionary conducted verification and eventually built a relatively high credibility semantic lexicon. (2) research product reviews corpus of product attributes (evaluated) extraction method. Product attribute extraction is a valuable research task, in order to improve corpus review the accuracy of product attribute extraction, this paper uses conditional random field model and maximum entropy model in the product attribute extraction task for comparative analysis. In addition, also on the part of speech, and other relevant characteristics of shallow syntactic selected for a detailed description, analysis add these syntactic characteristics on the impact of product attributes feature extraction. (3) explore the text sentiment analysis of product attributes interdisciplinary transplant method. Text corpus for the current lack of emotional status, is proposed based on the evaluation of active learning objects interdisciplinary transplantation. This model from the field of electronic products will migrate to the automotive sector, the experimental results show that the proposed method based on active learning in the field of transplantation emotions played a good role. (4) design and implementation of a fine-grained sentiment analysis system, the main features include the evaluation object extraction, polarity word recognition, evaluation object polarity judgment, emotional comprehensive analysis of the text of this article the research results.

Related Dissertations

  1. Based on conditional random RNA secondary structure prediction algorithm,R346
  2. Based on terms of joint decoding of Chinese word segmentation,TP391.1
  3. Application of Controlled-Release Fertilizer on Strawberry (Fragaria Ananassa Duch.) in Greenhouse,S628
  4. Text- based English blog space analysis of emotion,TP391.1
  5. Chinese named entity recognition and disambiguation of,TP391.1
  6. Research on Sentiment Analysis of Product Reviews Based on Domain Ontology and Conditional Random Fields,G353.1
  7. Domain-specific Chinese term extraction,TP391.1
  8. Fast and accurate method of structured machine learning,TP181
  9. Localization and Mapping for Mobile Robot in Large-scale Unstructured Environment,TP242
  10. Research on Human Motion Modeling and Pose Estimation from Monocular Videos,TP391.41
  11. Study on Matrix-Based (Humic Acid) SRNFs and Their Sweet Potato Biological and Rhizosphere Soil Environmental Effects,S531
  12. Characters of Nitrogen Nutrition in Ground Cover Rice Production System,S511
  13. Modern Chinese verb - multi-angle inspection and automatic identification,H146.312.440
  14. Distributed video coding side information fusion and reconstruction,TN919.81
  15. Adaptive Evolution of HIV-1 and A/H1N1 and Evaluation of Bootscan Analysis of HIV-1 Recombination,R373
  16. Comments mining semantic information extraction,TP311.13
  17. CRFs-Based Chinese Named Enitity Recognition with Improved Tag Set,TP391.1
  18. Based on Conditional Random stratified research paper information extraction research,TP391.1
  19. Research on Content-based Shot Classification in News Video,TP391.41
  20. Research on the Automatic Term Extraction in the Area of Information Science,G250

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Text Processing
© 2012 www.DissertationTopic.Net  Mobile