Hotel Recommendation System Based On Hadoop

Today,with the ever-changing Internet technology,the overload problem is becoming more and more prominent.The recommendation system is the optimal solution to this problem.It can quickly match the content they need in the massive information based on the user's interest.In reality,because of the large amount of relevant data between items and objects,people and people,the recommendation system based on a single machine has difficulty in recommending calculations for a large amount of information data because of limited performance.A number of solutions have emerged for the scalability of recommended systems,all based on the distributed open source software architecture Apache Hadoop.This paper studies the HDFS distributed architecture and discusses the MapReduce programming model.Combined with the collaborative filtering recommendation algorithm widely used in recent years,this paper proposes some distributed parallel algorithms based on MapReduce programming theory,and the Mahout recommended algorithm framework is Encapsulation of the recommended algorithm MapReduce programming.The details are as follows:1?This paper studies the HDFS distributed architecture and discusses the MapReduce programming model,and then combines with the widely used collaborative filtering recommendation algorithms and systems in recent years.MapReduce parallel programming is the main programming method of the system,and the complex task of the algorithm is decomposed,making it a small task for multiple MapReduce jobs.In order to enable distributed parallel processing to be implemented on the Hadoop framework,different types of collaborative filtering are used.The recommended algorithm was verified,which confirmed that the design has strong scalability and good parallelism in the cluster.2?The data source structure of the system is studied.The recommended module of the system adopts the data of the hotel security information system,and then deeply analyzes the internal structure of the data to extract and extract valuable data.3?The use of a variety of open source frameworks,including Hadoop,Storm,Sqoop,Mahout and other open source frameworks or algorithm frameworks.Since MapReduce is used to implement the hotel collaborative filtering algorithm,it is more helpful to explain the essence of the recommendation algorithm.Therefore,this paper focuses on how to use MapReduce to implement the hotel recommendation algorithm.At the system implementation level,the Mahout open source recommendation algorithm framework is adopted,which is a package for the MapReduce collaborative filtering recommendation algorithm,which is more conducive to development.4?The implementation of the basic information system was studied.Since the final result of the recommendation module needs to be presented to the customer,it is necessary to simply study how to use the framework of the Java open source community to develop the web application;the hotel recommendation module and the web system are independent.This article focuses on the hotel recommendation module,a brief introduction to the hotel web information system.