- 1 min

Sentimental is an easy-to-use Python library for doing sentiment analysis. That is, it uses machine learning to try to label sentences or paragraphs as negative, neutral or positive depending on the sentiment being expressed.

What’s the magic sauce?

Nothing! Absolutely nothing! It’s all built upon standard algorithms and methods for doing sentiment analysis.

Sentimental uses Logistic Regression on a tf-idf matrix of n-grams. Then the model is evaluated using k-fold crossvalidation.

The difference between Sentimental and other libraries is its ease of use. Once you have a training set it takes minutes to set up a program that uses Sentimental and start to label sentences.

As with all machine learning, the training data is the most important part and that is what determines the accuracy.

Getting started with Sentimental

Sentimental is released open source under Apache 2.0 and is available on my GitHub page.

Erik Gärtner

Erik Gärtner

Research Scientist at RADiCAL