We see how vector space models help to identify similar sentences with different words and also helps capture dependency between the words.
Similarity Metrics used :
- Euclidean distance
- Cosine Similarity
- Manipulated word vectors to find the relationship between 2 words based on a given relationship. Eg : King - Man + Woman = Queen
- Word Embeddings Manipulation (Vector Space Model)
- Principal Component Analysis implementation
- Finding word analogies using different similarity metrics
- Python 3.6.10
- Numpy 1.18.4
- Pandas 1.0.5
- Matplotlib 3.2.1
- NLTK 3.5
- Scikit-learn 0.23.1
Model accuracy : 92.0