- Using bag of words to convert the useful information (word vector)
- Calculated the cosine similarity (needs to be close to 1)
- Depending on the cosine similarity, I found the top 10 different recommendations
- Trained classification models on my data
- Calculated metrics: Accuracy, Precision, and Recall
- Tested the model on the data collected
- Google Colab
- Python Libraries:
- NLP Libraries: spacy, and gensim
- Matplotlib, seaborn, sklearn, pandas, etc
- When I came across some issues I looked through different youtube videos to see all the different approaches people tried.
- I was also able to find more effective ways to debug my code.
- I was able to make a basic recommender system where I was able to give the 10 recommended posts
- Was able to use a visual representation to analyse the results
- Clean the data so that I can use it in my recommender system
- Became more comfortable with using google colab
- Cleaned the data so that it can be used for my model (data pre-processing)
- Was able to understand and use cosine similarity to define my recommender system
- Trained and tested my model and data
- I had a hard time making the recommender system so I looked at different videos to figure out the different ideas that I could use.
- The cosine similarity idea was a little difficult for me to conceptualize, but eventually I was able to understand it by watching different videos.