Name		Name	Last commit message	Last commit date
parent directory ..
nearest-neighbors-features-and-metrics_blank		nearest-neighbors-features-and-metrics_blank
README.md		README.md
nearest-neighbors-features-and-metrics_blank.html		nearest-neighbors-features-and-metrics_blank.html
nearest-neighbors-features-and-metrics_blank.ipynb		nearest-neighbors-features-and-metrics_blank.ipynb
nearest-neighbors-features-and-metrics_blank.py		nearest-neighbors-features-and-metrics_blank.py

README.md

Goal

Gain intuition for different notions of similarity and practice finding similar documents.
Explore the tradeoffs with representing documents using raw word counts and TF-IDF
Explore the behavior of different distance metrics by looking at the Wikipedia pages most similar to President Obama’s page.

File Description

.zip file is data file.
- people_wiki.csv.zip (unzip people_wiki.csv) consists of 59,071 pages and 3 features. URL name text
.json files
- people_wiki_map_index_to_word.json
.npz files
- people_wiki_tf_idf.npz
- people_wiki_word_count.npz
description files
- .ipynb file is the solution of Week 2 program assignment 1
  - nearest-neighbors-features-and-metrics_blank.ipynb
- .html file is the html version of .ipynb file.
  - nearest-neighbors-features-and-metrics_blank.html
- .py
  - nearest-neighbors-features-and-metrics_blank.py
- file
  - nearest-neighbors-features-and-metrics_blank

Snapshot

Recommend open md file inside a file
open .html file via brower for quick look.

Algorithm

K-NN with word count and tf-idf
metrics differ from Euclidean and Cosine

Implement in details

Extract word count vectors
Find nearest neighbors using word count vectors
Interpreting the nearest neighbors
Extract the TF-IDF vectors
Find nearest neighbors using TF-IDF vectors
Choosing metrics
Problem with cosine distances: tweets vs. long articles