Sentiment analysis for an Ecommerce Platform
The commercial department of an ecommerce company wants to analyze their website's comments and reviews to improve their users' experience.
- Industry
- Retail
- Solution
- Advanced Analytics & AI
01 / 03
The challenge
The commercial department of a company in the e-commerce industry constantly receives comments and reviews from their customers on its website.
- The company wants to analyze those comments and reviews to obtain valuable information about the opinion of their customers about their products and services.
02 / 03
What Tekne did
Objective
Sentiment analysis is a natural language processing technique that allows the evaluation of text content in terms of emotions and opinions, and can be used in various fields, including e-commerce.
In this case, TensorFlow, which is a Machine Learning platform developed by Google, was used to build a sentiment analysis model that automatically provides a positive, negative or neutral probability for comments.
Process and architecture
- The model is trained on a pre-tagged data set, which includes comments and reviews with their respective rating of positive, neutral, or negative.
- Once trained, the model can be used to analyze new comments and reviews and rank them based on their sentiment.
03 / 03
Results
The company used the information obtained from the sentiment analysis model to:
- Improve their products and services
- Serve their customers more efficiently
- Generate reports and statistics to evaluate the performance of their products
