Linkedin Similarity with Word Embeddings

• I was one of the 10 students in my university to receive a honors grant to carry out an independent research.
• Wrote an undergraduate NLP research article comparing different text similarity method on Linkedin Data
• Implement and analyzed performances of different NLP methods such as Word2Vec and GloVE on text similarity.

Paper Description

In this paper, we will primarily be comparing the standard bag- of-words approach with a popular word embedding by Stanford

called GloVe. We will be implementing these two embeddings with a popular text similarity method, namely, cosine similarity. We also explore the possibility of using text similarity to calculate the similarity between LinkedIn profiles. We will be running the two different approaches to compare the similarity of different LinkedIn profiles. We will then visualize the result using a heatmap to analyze our result.

Link to paper

• Paper - Undergraduate Research Journal Paper

Credits

• Dr. Kevin Lee

---