This project is a graduation Thesis developed by George Smyrnaios and Giannis Moschonas.
The project has three parts:
- Source code: https://github.com/csdRepo/OpinionMining
- GUI (HTML5 using bootstrap): https://github.com/csdRepo/OpinionMiningGUI
- Text (LATEX): https://github.com/csdRepo/OpinionMiningLatex
Heraklion, 2015, Univercity of Crete
Opinion Mining is a field in Computer Science where a better association between Natural Language and Computers is attempted (since it is part of the wider Natural Language Processing) as well as an attempt to mine semantic information from writing. More precisely, in the wider field of "Natural Language Processing", "Machine Translation", "Named Entity Recognition and Disambigution", "Sentiment Analysis" etc. are included. In this Thesis, a better approach to information extraction from plain texts is attempted. The texts contain commentaries of citizens on consultation of Laws issued by the Greek Government. Extraction of proposals and counterproposals that the writers of these texts make, as well as of the arguments they formulate, is attempted. Finally, the general opinion of the writer is extracted, which is summarised by a "Positive" or "Negative" characterisation, depending on the writer's view, which comes of the while text. Mining of the above data is done entirely by analysing the texts through a three-stage analysis (which will be explained in detail in the following chapters of this Thesis) and by utilising techniques that are included in the wider spectrum of NLP (Information Retrieval, Part-of-Speech Tagging, Sentiment Analysis etc.). The results show that we can create realistic methods that can extract this type of information. This matter troubles the Scientific community because these methods can be put into use in a variety of other fields, apart from the field of Computer Science (e.g. Journalism, Politics etc.).