Maitreyee Tewari

Recognizing Trust in Natural Language in Amazon's Online Reviews: Extracting Ethotic Statements About Amazon's Service from a Corpus of Amazon's Customer Reviews Maitreyee Tewari

The project?s motivation is to make internet a credible world. This project is a first step towards automating the process of trust extraction. The project involves analysis, extraction and refinement of ethotic data available on the internet. The project was initiated with gathering of data from different sources such as online libraries or the website. Then the data was cleaned (removed stop words and punctuations), annotated and classifier was built. The entire project went through several iterations to produce a reliable solution. This project has used several natural language processing techniques such as argument mining, sentiment analysis and machine learning algorithms such as Naive Bayes and Maxent Models which are discussed in detail in the thesis. The software involved were Python and its toolkits for language processing such as Natural language Toolkit (NLTK) and TextBlob. The contribution of this project was: a rich data was collected and annotated and two classifiers were built which could extract ethotic statements and classify them.