Hadoop realizes shopping mall recommendation system

1, shopping mall: is a single merchant, many buyers of the http://www.aliyun.com/zixun/aggregation/36896.html "> Mall system." Database is MySQL, language java.

2,sqoop1.9.33: Exchanging data in MySQL and Hadoop.

3,hadoop2.2.0: This is a pseudo distribution pattern for practice.

4, complete the content: People like the product also like, the same shopping preferences friends recommend.

Steps：

1, through Sqoop from MySQL in the "User Collection Products" (here with the user's collection of merchandise information table as a recommendation system business basis, business basis can be very complex. This article mainly introduces the principle of the recommendation system, so it is recommended to import the table data into HDFs.

2, using MapReduce to implement the recommendation algorithm.

3, the results of the recommendation system will be written back to MySQL via Sqoop.

4,java Mall through the recommendation system data to achieve < like the product people also like, the same shopping preferences friends recommend. > Two features.

Achieve：

1, the recommended system data sources:

The left is the user, the right side is the merchandise. Each collection of users will generate a product such information,< like the product people also like, the same shopping preferences friends recommend. The data source for > is this table.

Sqoop Import data, here is the sqoop1.9.33. sqoop1.9.33 information is very small, there will be some errors, can not be sent to the search of my mailbox keepmovingzx@163.com.

Create link information

It's easier.

Create Job

You can fill in the information.

Import data Execute start job--jid The ID returned after the successful creation

Import successful data

2,eclipse Development MapReduce Program

shopxxproductrecommend< people like this product also like >

The whole project is divided into two parts, one, the user to the commodity group, two, to find the same present matrix of goods.

One, the 1th big step Data group the items for input parameters

Output parameters:

Second, the first step of the output data for the input to find the same present matrix of goods

Output data

The first column is the current product, the second one is similar to it, and the third is the similarity rate (the higher the more similar).

The whole process is over.