An Evaluation of Word Frequency Techniques for Text Summarization Using Sentiment Analysis Approach

Digital data has become an important aspect of machine learning and is present in huge volumes on the internet. To use this data efficiently, data handling and processing techniques are required to filter out information from documents and store them. An application of natural language processing, which helps in handling volumes of data, is text summarization. Text summarization helps in condensing documents, and extract the important facts represented in it. There are two techniques in text summarization: abstractive and extractive summarization. Extractive Summarization extracts keywords from the document and combines them to provide a semantically incorrect summary, whereas, Abstractive Summarization produces a semantically correct summary of the text. In this paper, we compare different techniques to identify low and medium frequency words in hotel reviews. We evaluate the techniques based on the correct identification of positive and negative words.