LEARN.md

Recommendation in E-commerce

• Personalized recommendation algorithm

TODO

• Movie recommender system
• Recommender System in Python 101
• Recommender System using Amazon Reviews
• Matrix factorization using the surprise library
• Links to real-time training + prediction models

Datasets

• Amazon Reviews

TFIDF

[1] Note that the tf-idf functionality in sklearn.feature_extraction.text can produce normalized vectors, in which case cosine_similarity is equivalent to linear_kernel, only slower.

from sklearn.feature_extraction.text import TfidfVectorizer
tfidf = TfidfVectorizer(stop_words='english')
tfidf_matrix = tfidf.fit_transform(X)

cosine_similarities = linear_kernel(tfidf_matrix[target_idx], tfidf_matrix).flatten()
related_docs_indices = cosine_similarities.argsort()[:-5:-1] # Meaning reverse the list, but take the last 5 only.
cosine_similarity[related_docs_indices]

• 1
• https://datascience.stackexchange.com/questions/62143/ts-ss-and-cosine-similarity-among-text-documents-using-tf-idf-in-python
• https://stackoverflow.com/questions/12118720/python-tf-idf-cosine-to-find-document-similarity/12128777#12128777