Research On Cloud Services Of Data Mining Based On Hadoop Big Data Platform

With the rapid development of network technology, mobile Internet technology, social networking, sensor technology, huge amounts of data constantly being generated.The world is entering the era of the IT era to DT. With the rapid growth in data size, the construction which medium-size enterprises (SMEs) create based on the traditional stand-alone data analysis system are obviously unsatisfactory in server utilization, storage, data mining, Unable to effectively solve many problems of SMEs in the value of information technology implementation process faced.Obtain valuable data analysis has become an important means for SMEs to enhance their competitiveness. This paper studies how to collaborate division, data collection, transmission, storage, processing and so on. How to build Hadoop big data platform to sovle the lack of traditional data warehouse in the mass data processing, storage and so on. supply data mining cloud services based on one of three modes of Saas (Software as a Service) mode.Rely on Hadoop’s ability to scale to break the traditional single-node data warehouse bottlenecks in data mining. The data mining cloud service uses MapReduce of Hadoop platform for parallel computing, excavation for Hadoop distributed file storage system(HDFS) in vast amounts of data. To further verify the stability of the platform, the use of fusion improved K-Means algorithm and improved Apriori algorithm on the platform more efficient mining the data which business concern.Experimental results show that the running of the improved K-Means algorithm and improved Apriori algorithm on Hadoop platform analysis of vast amounts of data, can significantly improve the accuracy and efficiency of data mining results. Therefore, data mining cloud services can effectively solve the problem of SMEs in the capital, talent, technology, which led to obtain valuable data more difficult problems, meet their individual, demand flexibility through a variety of algorithm model. Currently help decision makers make better decisions has become a cloud of data mining technology in the field of new topics. In this paper, by analyzing needs of SMEs and data mining the value of cloud services, providing valuable information for SMEs.The main work and achievements are accomplished:1、For Hadoop, big data platform, data mining, development status of cloud services were studied and analyzed, discussed the big data platform structures and data mining cloud services discussed to provide data mining cloud-based Hadoop BigData platform of significance, We completed the theoretical background research this topic.2、Combined with theoretical research, with reference to a large number of domestic and foreign research results, taking fully into account the value of the data requirements of small and medium enterprises of the actual situation, clear the research purposes of data mining cloud service, from the user’s needs to establish a big data platform based on Hadoop.3、The use of virtualization technology and related technology of bigdata to build big data platform based on Hadoop that provides data mining cloud services to SMEs.4、Big data platform to run on improved Apriori parallel algorithm and improved k-Means algorithm code jar package to obtain accurate data mining results.5、Achieve real-time, fast, convenient and accurate analysis of data mining and solve the difficulties of traditional data mining can not handle large amounts of data, the cost savings for SMEs enjoy high-value data analysis, reducing the value of the technical requirements for SMEs to obtain data for a variety of industries massive data mining, data value for SMEs and make the right decisions has important significance.