At its core, the task involves taking customer review data for a certain product as input and outputting a summary which includes the number of positive and negative sentiment-bearing phrases for the features of the product (screen, size, camera, etc.).
This project is broken into 7 sections:
- Analyse the task.
- Preprocess the data.
- Extract relevant information.
- Opinion Words Extraction.
- Apply a relevant sentiment analysis algorithm and experiment with a model extension.
- Report evaluation results.
- Print summaries.
Relevant evaluation metrics are also reported throughout the different steps of my implementation indicated by blue boxes along with a dedicated final section.