Aspect Based Sentiment Analysis on E-Commerce Reviews

This blog is an introduction to the series of ABSA. Followed by the blogs regarding the various steps involved. The basic idea of this blog is to give intuition about the Aspect Based Sentiment Analsysis (referred as ABSA) and why it’s gaining popularity among organizations.

Image Credit: Nitesh Tripathi , Aspect-Based Sentiment Analysis in Product Reviews: Unsupervised Way

What is Sentiment Analysis — Application, Benefits, and Limitations

Sentiment Analysis is a computational study or technique to distinguish positive and negative opinions from textual data pro grammatically. Many cutting-edge technologies like Natural Language Processing (NLP), Machine Learning (ML), Text Processing, and Deep Learning (DL) are being used nowadays to automate sentiment analysis. It allows to score within a quantified range — the positive, negative or neutral sentiment from a piece of text, with less human effort.

Sentiment analysis may be applied in multiple areas such as customer feedback, movie or product reviews, and political comments. Large enterprises perform sentiment analysis to analyze public opinion, conduct market research, monitor brand, product reputation, and understand customer experiences.

Why Just Sentiment Analysis not enough?

While sentiment analysis can help identify the sentiment behind an opinion or statement, there might be several aspects that have triggered the identified sentiment. For instance, when analyzing reviews, it is easier to comprehend positive reviews than negative ones. Also, it requires determining the intended ‘aspect’ of the review that has generated a negative opinion.

Like people buy different items such as food, clothing, electronic items etc though e-commerce websites. And they write about their experiences there. For instance, a person who bought clothing from an e-commerce giant wrote “Very nice but the length is a little small but the shirt is very nice and very stylish.” (This example is from Amazon). He has raised his concern about the shirt length. And he is happy about the style shirt has brought to his life. But the company who sells that shirt should know that the shirt has fitting issues. Here comes the Aspect Based Sentiment Analysis to the rescue!

Understanding Aspect Based Sentiment Analysis

Aspect Based Sentiment Analysis (ABSA) is a technique that takes into consideration the terms related to the aspects and identifies the sentiment associated with each aspect. ABSA model requires aspect categories and its corresponding aspect terms to extract sentiment for each aspect from the text corpus. One can create a domain-specific model for a specific implementation.

Typical ABSA requires labeled data containing aspect terms and aspect categories for each statement along with its sentiment score. However, it can be solved using the unsupervised approach without having labeled data and a list of aspect terms. For example, what was the overall experience of customers who bought the shirt in terms of  the styling, fitting, cost?

Image Credit: Aspect-based Sentiment Analysis — Everything You Wanted to Know!

A business needs to identify the aspects of the product/service that attract more customers and/or keep away people from using the product/service. ABSA identifies sentiment for each aspect category i.e. styling, fitting, cost. It helps business to track how end-users sentiment changes toward specific features and attributes of a service or product

Implementation Details

We at Paxcel Labs have implemented ABSA using the unsupervised approach.The complete project is divided into three modules.

  • Module 1: Cleaning and finding aspect terms. Aspect terms are basically, the phrases in the review written by the customers that describe the product and sentiment related to it.
  • Module 2: Finding/Predicting Aspect Categories from aspect terms. Categories are the bins that relate the extracted phrase to the property of the product.
  • Module 3: Finding sentiment polarity of the aspect terms/reviews. From the extracted phrase find whether the customer was satisfied with it(Pos:Positive) or was it an unsatisfactory experience(Neg:Negative).

Summary

In this blog we got understanding about the Aspect Based Sentiment Analysis (ABSA) and how it is so important for an organization to get an understanding of how they are performing in the market they are targeting. In the next blogs we will discuss the steps involved in the ABSA. Those steps are developed in such a way that they do not get restricted to the e-commerce domain, but  can be used anywhere according to the problem statement.