suggestion mining from online user reviews using information retrieval methods - دانشکده فنی و مهندسی
suggestion mining from online user reviews using information retrieval methods
نوع: Type: thesis
مقطع: Segment: masters
عنوان: Title: suggestion mining from online user reviews using information retrieval methods
ارائه دهنده: Provider: zahra hadizadeh
اساتید راهنما: Supervisors: Dr. Moharram Mansoori zadeh
اساتید مشاور: Advisory Professors: Dr. Mir Hossein Dezfolian
اساتید ممتحن یا داور: Examining professors or referees: Dr.Reza Mohammadi, Dr. Morteza Yousef Sanati
زمان و تاریخ ارائه: Time and date of presentation: 2021
مکان ارائه: Place of presentation: Department Engineering
چکیده: Abstract: With the increasing growth of the web and the possibility of interaction between users and the expression of opinions on the web, comment mining has become a popular area of research in the field of natural language processing. User feedback largely expresses positive and negative feelings about a particular entity, and in addition to these, they make suggestions and suggestions for improving organizations and other users' decisions available to the public. The purpose of this study is to use information retrieval techniques to automatically categorize opinions. With the help of information retrieval, sentences of a text that are unstructured are converted into vector numbers. Then, with classification algorithms, they are automatically divided into proposed and non-proposed categories. Given the very limited volume of related work, proposal extraction can be considered as a new research issue. Therefore, in this dissertation, distance measurement approaches, perceptron multilayer neural network, support vector machine, regression logistics, and torsional neural network with TF IDF, bow and word2vec vectors, and keyword extraction were performed. The proposed method is presented on the Semeval2019 data set, Task9, for extracting suggestions from the text of online comments. The results show that the f1-Score has improved compared to the previous work and has reached 0.87
فایل: ّFile: Download فایل