GitLab

Created by: J3FALL

Impement experimental pipeline with simple NLP-model (or even NaiveBayesClassifier) for text classification/sentiment analysis:

• New InputData format with generator-based logic (for texts batch-loading)
• New simple model (for instance, CNN) or adoptation the existing models that are suitable for text classification
• Maybe bag-of-words, word-2-vec model, etc. for text embeddings
• Several text preprocessing methods (stemming, lemmatization, filtering, etc.)
• Example of dataset where pipeline can be applied (spam classification for instance as the simpliest case)
• Example of FEDOT pipeline that can solve the given task