Multilingual Sentiment Analysis on Product Reviews

This is a bert-base-multilingual-uncased model finetuned for sentiment analysis on product reviews in six languages:

• English
• Dutch
• German
• French
• Spanish
• Italian

It predicts the sentiment of the review as a number of stars (between 1 and 5).

This model is intended for direct use as a sentiment analysis model for product reviews in any of the six languages above or for further finetuning on related sentiment analysis tasks.

Requirements

For the purpose of this project, I've used the Amazon Review Scraper Extension on a review page.

Here's an example of the input file you should upload on the app

Key Advantages of this Model for Sentiment Analysis

1. Accuracy of the Sentiment's Magnitude. Compared to other models, such as cardiffnlp/twitter-roberta-base-sentiment-latest, which return textual classifications (negative, neutral, positive), the multilingual-uncased model provides scores ranging from 1 to 5 to indicate the sentiment's magnitude. This approach helps disambiguate text classification for improved accuracy in the output (e.g., does a neutral review correspond to a score of 2 or 3?).

2. Fine-tuned to Multilanguage text. This model is able to capture the linguistic nuances from different langauges. Accuracy could be improved but still it does the job.

Useful Links

1. Hop on the Multilingual Sentiment Analysis Review app
2. Find more at my SEO Toolstation