Name: SWE546_Assignment4
Owner: Bogazici-SWE546-Spring2016
Description: MovieLens with Stochastic Gradient Descent
Created: 2016-04-06 16:07:39.0
Updated: 2016-11-19 22:32:44.0
Pushed: 2016-06-04 11:00:57.0
Size: 6015
Language: null
GitHub Committers
| User | Most Recent Commit | # Commits |
|---|
Other Committers
| User | Most Recent Commit | # Commits |
|---|
MovieLens Recommendation System
http://grouplens.org/datasets/movielens/
Using the MovieLens 1M dataset, and using only the rankings, implement a recommendation sytem using SGD.
import pandas as pd
rnames = ['user_id', 'movie_id', 'rating', 'timestamp'] ratings = pd.read_table('data/ml-1m/ratings.dat', sep='::', header=None, names=rnames)
The output of your program must be the highest estimated rankings for the first 10 movies of the first 20 users
Modify the program for mini batch and classical gradient descent.
Compare the running times and solution quality as a function of the batch size. SGD is when batchsize = 1, Gradient descent is batchsize = Number of observed entries in the matrix.
Submit a short report with the plots of your results.