Research On The Improvement Of Collaborative Filtering Recommendation Algorithm Under Big Data Environment

With the development of the global economy, technology innovation becomes animportant factor in promoting the development of productivity. How can technology betransformed into productivity quickly is an important research topic throughout the world.With the support of information methods, technological innovation platform integrates thestrength of industry-university-research, promotes the realization of science and technologyachievements as well as the development of industrialized and strategic new industry. Italso facilitates the transformation from technical innovation to technology.Due to the promotion of the application of the technological innovation platform,technological innovation platform projects and the supply and demand of users increaserapidly, the explosive growth of the amount of data appears, it’s very difficult for suppliersand demanders to find the right technology docking project in the massive amounts ofinformation quickly. The technical recommendation system can make personalizedrecommendation for users to improve the success rate of the technological docking.However, rapid growth in the amount of technological innovation platforms and users today,the massive complex, multi-source heterogeneous data, natural language expressive, aswell as potential user needs become a bottleneck of the technical recommendation systemto produce accurate recommendation results quickly and efficiently. Under the environmentof big data, this paper integrates ontology with the modern intelligent recommendationalgorithm and big data solutions to improve the accuracy and speed of technicalrecommendation.This paper first analyzes the characteristics, objectives and the current research statusof the technological innovation platform and technology recommendation system. it alsoanalyzes the needs of users, and finds the existing problems in technical recommendationsystem under the environment of big data. Combined with the characteristics of big data,on the basis of analysis of the current research status of the technology recommendationsystem, this paper selects the collaborative filtering recommendation algorithm for itslower data requirements and more successful application. The collaborative filteringrecommendation algorithm is difficult in handling the natural language expressive andpotential user needs under the environment of big data, which would affect the recommendation accuracy. In order to solve this defect, combined with the semanticsimilarity into the traditional collaborative filtering algorithm. This paper introducesontology to improve technical recommendation model, improves the domain ontologyknowledge base(DOKB), on one hand, it improves the accuracy of recommendation, tapsthe dynamic and potential needs of users, on the other hand, transforms the user needsexpressed by the text into location information of the DOKB, therefore improves the speedof recommendation. In order to test the effectiveness of the improved algorithm, exampletest results show that the improved algorithm can improve recommendation efficiency andrecommendation quality to some extent.However for the massive complex big data, the introduced ontology optimizedcollaborative filtering algorithm is lacking the speed of recommendation. Through theanalysis of the existing big data solutions, this paper selects MapReduce which is relativelysimple and convenient to further improve the algorithm to increase the speed ofrecommendation. Example test results show that the speed of the collaborative filteringalgorithm based on ontology optimized by MapReduce is further improved.In view of the big data characteristics of the technological innovation platform, thispaper improves the collaborative filtering algorithm based on ontology. Combined with theprocessing technology of MapReduce, it has a certain value for the future development andapplication of the technological innovation platform and technology recommendationsystem.