Music Recommender System Using Pyspark

提供:
Coursera Project Network
このガイド付きプロジェクトでは、次のことを行います。

Learn how to setup the google colab for distributed data processing

Learn how aggregate a pyspark dataframe to have the data needed for our machine learning model

Learn how to use StringIndexer to convert a String (categorical) column into Unique Integral column

Learn how to create ALS model for Recommender System

Clock1 hour
Intermediate中級
Cloudダウンロード不要
Video分割画面ビデオ
Comment Dots英語
Laptopデスクトップのみ

Nowadays, recommender systems are everywhere. for example, Amazon uses recommender systems to suggest some products that you might be interested in based on the products you've bought earlier. Or Spotify will suggest new tracks based on the songs you use to listen to every day. Most of these recommender systems use some algorithms which are based on Matrix factorization such as NMF( NON NEGATIVE MATRIX FACTORIZATION) or ALS (Alternating Least Square). So in this Project, we are going to use ALS Algorithm to create a Music Recommender system to suggest new tracks to different users based upon the songs they've been listening to. As a very important prerequisite of this course, I suggest you study a little bit about ALS Algorithm because in this course we will not cover any theoretical concepts. Note: This project works best for learners who are based in the North America region. We’re currently working on providing the same experience in other regions.

あなたが開発するスキル

Programming ModelAlgorithmsAlgorithm TrainingPySpark

ステップバイステップで学習します

ワークエリアを使用した分割画面で再生するビデオでは、講師がこれらの手順を説明します。

  1. Prepare the Google Colab for distributed data processing

  2. Mounting our Google Drive into Google Colab environment

  3. Importing csv file of our Dataset (4 Gb) into pySpark dataframe

  4. Dropping some useless columns and nan Values in our dataframe

  5. Performing an Aggregation to prepare the data

  6. Learn how to use StringIndexer to convert a String (categorical) column into Unique Integral column

  7. Creating ALS model for Recommender System

ガイド付きプロジェクトの仕組み

ワークスペースは、ブラウザに完全にロードされたクラウドデスクトップですので、ダウンロードは不要です

分割画面のビデオで、講師が手順ごとにガイドします

よくある質問

よくある質問

さらに質問がある場合は、受講者向けヘルプセンターにアクセスしてください。