Giter VIP home page Giter VIP logo

movietaster-open's Introduction

MovieTaster-Open

A movie recommend project based on Item2vec.

Reference:

  • Barkan, Oren, and Noam Koenigstein. "Item2vec: neural item embedding for collaborative filtering." Machine Learning for Signal Processing (MLSP), 2016 IEEE 26th International Workshop on. IEEE, 2016.
  • JayveeHe, https://github.com/JayveeHe/MusicTaster. Github.

Demo>

More details for this project, plese refer to blog> or zhihu>

Project Struct

  • datas: to store corpus
  • models: to store item vectors for movies
  • utils: a collection of tool functions

Usage

  1. Compile fasttext under root of the project

  2. Process corpus, the data file would be generated under ./datas/.

$ cd datas && tar -xzvf corpus.tar.gz
$ cd ..
$ python utils/process.py
  1. Train model by running fasttext. Please refer to ./run.sh for other configuration.
$ ./fasttext skipgram -input ./datas/doulist_0804_09.movie_id -output ./models/fasttext_model_0804_09_skipgram -minCount 5 -epoch 50 -neg 100
  1. Here's a result from the default configuration.
$ python utils/eval.py
movie_name_id_dict is 130206 from [./datas/doulist_0804_09.json]
movie_name_id_dict is 130206 from [./datas/doulist_0804_09.json]
movie_id_name_dict is 130206 from [./datas/doulist_0804_09.json]
[2017-08-18 07:52:53,579][pid:22831] eval.topk_like: INFO: [1210]小时代 top 5 likes:
[2017-08-18 07:52:54,149][pid:22831] eval.topk_like: INFO: [2396]小时代2：青木时代 0.820323
[2017-08-18 07:52:54,149][pid:22831] eval.topk_like: INFO: [3387]分手合约 0.606063
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [4087]不二神探 0.604207
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [3839]天台爱情 0.601879
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [821]致我们终将逝去的青春 0.600836
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [144]倩女幽魂 top 5 likes:
[2017-08-18 07:52:54,690][pid:22831] eval.topk_like: INFO: [2719]倩女幽魂3：道道道 0.685621
[2017-08-18 07:52:54,690][pid:22831] eval.topk_like: INFO: [1812]倩女幽魂2：人间道 0.681594
[2017-08-18 07:52:54,691][pid:22831] eval.topk_like: INFO: [466]胭脂扣 0.678032
[2017-08-18 07:52:54,691][pid:22831] eval.topk_like: INFO: [261]青蛇 0.671541
[2017-08-18 07:52:54,692][pid:22831] eval.topk_like: INFO: [156]东邪西毒 0.664057
[2017-08-18 07:52:54,692][pid:22831] eval.topk_like: INFO: [2602]悟空传 top 5 likes:
[2017-08-18 07:52:55,253][pid:22831] eval.topk_like: INFO: [5648]闪光少女 0.671337
[2017-08-18 07:52:55,253][pid:22831] eval.topk_like: INFO: [2189]绣春刀II：修罗战场 0.646861
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [8297]逆时营救 0.634753
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [12571]京城81号Ⅱ 0.625549
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [10545]父子雄兵 0.623032
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [56]美国往事 top 5 likes:
[2017-08-18 07:52:55,789][pid:22831] eval.topk_like: INFO: [18]天堂电影院 0.756449
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [14]辛德勒的名单 0.737502
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [17]教父 0.735216
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [69]闻香识女人 0.734119
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [42]西西里的美丽传说 0.732334
[2017-08-18 07:52:55,791][pid:22831] eval.topk_like: INFO: [2644]战狼2 top 5 likes:
[2017-08-18 07:52:56,346][pid:22831] eval.topk_like: INFO: [1456]大护法 0.612700
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [2107]战狼 0.580637
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [2602]悟空传 0.580365
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [9941]建军大业 0.575422
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [19040]阿唐奇遇 0.573613

MovieTaster-Open

使用Item2Vec做电影推荐

参考:

  • Barkan, Oren, and Noam Koenigstein. "Item2vec: neural item embedding for collaborative filtering." Machine Learning for Signal Processing (MLSP), 2016 IEEE 26th International Workshop on. IEEE, 2016.
  • JayveeHe, https://github.com/JayveeHe/MusicTaster. Github.

Demo>

原理详情请参考>

目录结构

  • datas: 存放语料
  • models: 存放电影向量
  • utils: 工具类，包括生成推荐电影列表等

使用说明

  1. 在根目录下载并编译fasttext

  2. 处理语料。结束后将在./datas/下生成fasttext所需要的格式

$ cd datas && tar -xzvf corpus.tar.gz
$ cd ..
$ python utils/process.py
  1. 参考./run.sh中生成词向量的参数配置，任选一个。
$ ./fasttext skipgram -input ./datas/doulist_0804_09.movie_id -output ./models/fasttext_model_0804_09_skipgram -minCount 5 -epoch 50 -neg 100
  1. 查看模型效果
$ python utils/eval.py
movie_name_id_dict is 130206 from [./datas/doulist_0804_09.json]
movie_name_id_dict is 130206 from [./datas/doulist_0804_09.json]
movie_id_name_dict is 130206 from [./datas/doulist_0804_09.json]
[2017-08-18 07:52:53,579][pid:22831] eval.topk_like: INFO: [1210]小时代 top 5 likes:
[2017-08-18 07:52:54,149][pid:22831] eval.topk_like: INFO: [2396]小时代2：青木时代 0.820323
[2017-08-18 07:52:54,149][pid:22831] eval.topk_like: INFO: [3387]分手合约 0.606063
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [4087]不二神探 0.604207
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [3839]天台爱情 0.601879
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [821]致我们终将逝去的青春 0.600836
[2017-08-18 07:52:54,150][pid:22831] eval.topk_like: INFO: [144]倩女幽魂 top 5 likes:
[2017-08-18 07:52:54,690][pid:22831] eval.topk_like: INFO: [2719]倩女幽魂3：道道道 0.685621
[2017-08-18 07:52:54,690][pid:22831] eval.topk_like: INFO: [1812]倩女幽魂2：人间道 0.681594
[2017-08-18 07:52:54,691][pid:22831] eval.topk_like: INFO: [466]胭脂扣 0.678032
[2017-08-18 07:52:54,691][pid:22831] eval.topk_like: INFO: [261]青蛇 0.671541
[2017-08-18 07:52:54,692][pid:22831] eval.topk_like: INFO: [156]东邪西毒 0.664057
[2017-08-18 07:52:54,692][pid:22831] eval.topk_like: INFO: [2602]悟空传 top 5 likes:
[2017-08-18 07:52:55,253][pid:22831] eval.topk_like: INFO: [5648]闪光少女 0.671337
[2017-08-18 07:52:55,253][pid:22831] eval.topk_like: INFO: [2189]绣春刀II：修罗战场 0.646861
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [8297]逆时营救 0.634753
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [12571]京城81号Ⅱ 0.625549
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [10545]父子雄兵 0.623032
[2017-08-18 07:52:55,254][pid:22831] eval.topk_like: INFO: [56]美国往事 top 5 likes:
[2017-08-18 07:52:55,789][pid:22831] eval.topk_like: INFO: [18]天堂电影院 0.756449
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [14]辛德勒的名单 0.737502
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [17]教父 0.735216
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [69]闻香识女人 0.734119
[2017-08-18 07:52:55,790][pid:22831] eval.topk_like: INFO: [42]西西里的美丽传说 0.732334
[2017-08-18 07:52:55,791][pid:22831] eval.topk_like: INFO: [2644]战狼2 top 5 likes:
[2017-08-18 07:52:56,346][pid:22831] eval.topk_like: INFO: [1456]大护法 0.612700
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [2107]战狼 0.580637
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [2602]悟空传 0.580365
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [9941]建军大业 0.575422
[2017-08-18 07:52:56,347][pid:22831] eval.topk_like: INFO: [19040]阿唐奇遇 0.573613

Recommend Projects

  • React photo React

    A declarative, efficient, and flexible JavaScript library for building user interfaces.

  • Vue.js photo Vue.js

    🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.

  • Typescript photo Typescript

    TypeScript is a superset of JavaScript that compiles to clean JavaScript output.

  • TensorFlow photo TensorFlow

    An Open Source Machine Learning Framework for Everyone

  • Django photo Django

    The Web framework for perfectionists with deadlines.

  • D3 photo D3

    Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

  • javascript

    JavaScript (JS) is a lightweight interpreted programming language with first-class functions.

  • web

    Some thing interesting about web. New door for the world.

  • server

    A server is a program made to process requests and deliver data to clients.

  • Machine learning

    Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.

  • Game

    Some thing interesting about game, make everyone happy.

Recommend Org

  • Facebook photo Facebook

    We are working to build community through open source technology. NB: members must have two-factor auth.

  • Microsoft photo Microsoft

    Open source projects and samples from Microsoft.

  • Google photo Google

    Google ❤️ Open Source for everyone.

  • D3 photo D3

    Data-Driven Documents codes.