This repopsistory contains various data analysis technique and subsequent models based on Amazon fine food review dataset
- Exploaratory data Analysis
- Recomendation System
- Review Classification - using Random Forest + word2vec
- Trained a word2vec model using Gensim package with the following parameter value ( Word vector dimensionality = 300, Minimum word count =40, Number of threads to run in parallel = 3, Context window size = 10 and downsampling rate for frequent words = 1e-3).
- Created Feature vector to be used by random forest model by averaging word embeddings of all words in the review.
- Accuracy : 0.895
