case_study : Search Algorithm Relevance

Data :

Data consists of samples from 5 algorithms in .csv format. The .csv files contain observations of searches (1 row = 1 search), the number of returned results and the rank of the clicked result. A rank of 0 means that no result was clicked.

It is important to get good search results on top while keeping searches without result (clicked result) on reasonable margins.

Task 1:

Evaluate the given data to determine the “best” algorithm. Choose and create metrics appropriate for the task. Report as many metrics as to share the thoughts and can find and describe differences in the outcome. Also, involves using descriptive statistics and visualizations to guide through the exploration.

Task 2:

Answer the following questions:

• How to determine whether two words or phrases are synonyms?
• In general, how do you find relationships to use for query expansion?
• What are the trade-offs between using click-through rate (CTR) and conversion rate as search success metrics?!

Concluding remarks on Natural Language Search and it's applications.