close

Вход

Забыли?

вход по аккаунту

?

MovieGEN: A Movie Recommendation System

код для вставкиСкачать
MovieGEN: A Movie
Recommendation System
Gaurangi, Eyrun, Nan
Outline
•Introduction
•Related work
•Implementation
•Results (Demo)
•Summary
Introduction
•Recommendation systems are special
types of expert systems
•Why necessary?
•MovieGEN
–Takes in : Personal particulars
–Asks questions
–Recommends movies
Background & Related Work
•Other movie recommendation
systems – Use movie ratings
•We use
–Machine learning – Support Vector
Machine (SVM)
–Cluster analysis – K Means Algorithm
Implementation
•Machine Learning based
Preference Prediction
Implementation (contd.)
•Data description
•Data formatting for SVM
regression
Implementation (contd.)
Implementation (contd.)
Machine Learning
Output Vector
•Movie Ranking
–Based on output
from SVM
Movie
XML
Chosen Movies
Movie Ranking Algorithm
SVM based
Machine Learning
Model
Ranked Movies
K Means Clustering
Movie Clusters
•Clustering
–K Means
•Question generator
Question Generator
Question
Feedback to
Questions
Recommended
Movies
Evaluation of
recommendatio
n
Implementation (contd.)
•Movie Database
– 300 movies
– Information from IMDB
<MOVIE>
<NAME>Titanic</NAME>
<GENRE>Romance</GENRE>
<GENRE>Drama</GENRE>
<STARRING>Leonardo DiCaprio</STARRING>
<STARRING>Kate Winslet</STARRING>
<DIRECTOR>James Cameron</DIRECTOR>
<YEAR>1990s</YEAR>
<AGE_GROUP>Adults</AGE_GROUP>
<GENDER>Both</GENDER>
<RATING>7.2</RATING>
<OSCAR>11</OSCAR>
</MOVIE>
Demonstration
Документ
Категория
Презентации
Просмотров
40
Размер файла
302 Кб
Теги
1/--страниц
Пожаловаться на содержимое документа