-
We implementetd the search feature using TF-IDF, a heuristic to determine and rank the best matches for a given input query.
-
To determine how good our search engine is, we developed 3 independent IR systems to compare. They are based on:
- Title
- Description
- Content
- Ted Talks
- Clickbait Videos and Podcasts
[Insert screenshots (or links?) here]
- We evaluated the success of our content based IR system in comparison to the other 2 IR systems. Metrics we used:
- Order Unaware: Precision (P), Recall (R), Accuracy (A), F Measuree (F1 score)
- Order Aware: Mean Reciprocal Rank (MRR), Average Precision (AP), Mean Average Precision (MAP), Cumulative Gain (CG), Discounted Cumulative Gain (DCG), Normalized Discounted Cumulative Gain (NDCG)
[Insert results of different evaluation metrics]