Analysis On The Idea Topics Of Open Innovation Community Based On LDA

Under the impact of economic globalization and the Internet wave,if companies want to maintain their own competitive advantages,they need to continuously innovate in technology or ideology.Open innovation communities have emerged as the times require.When internal innovation resources cannot meet the challenges,more and more companies have begun to change the innovation model,from traditional closed innovation to open innovation,and have gradually launched their own open innovation communities to widely absorb and efficiently use external creative resources,enhancing market competitiveness.However,in such an era of information explosion,the open innovation community as a virtual platform can easily face the problem of information overload.Although many companies have established their own open innovation communities,there are problems such as low participation and out-of-control innovation processes which cause communities have not yielded the desired benefits and returns.Therefore,companies providing communities as a platform to attract users to provide ideas is only a foundation.Capturing the topics discussed by users in the community,scientific and effective management of the community can make it really work for the enterprise.This paper combs related literature of the open innovation community and the topic model,then combined with related theories and methods such as cognitive load theory,social impact theory,and sentiment analysis,applying topic model to the open innovation community.The Salesforce community is selected as the research object.The idea content and other data are crawled for grouping and preprocessing.The latent Dirichlet Allocation(LDA)method,which is widely used in the topic identification field,is used to extract and analyze the idea topics in the community.Furthermore,using Stata,correlation analysis and negative binomial regression analysis were performed on idea topics,idea emotional tendencies,idea content length,and number of idea votes to explore the influence of idea topics on idea popularity.Finally,the results of topic extraction and regression are analyzed to obtain research conclusions,and the limitations of this article and future research prospects are summarized.The innovation of this research is mainly reflected in two aspects.One is to apply LDA to the open innovation community and expand its application field;the other is to regard the idea content itself instead of other characteristics as research subjects,and explore the influence of idea topics on idea popularity.Researching the idea content in the community,mining topics of user ideas and analyzing their changes over time,will help companies grasp the community content as a whole,understand and screen out valuable ideas in a targeted manner,promote product development and better management of users.This research provides a scientific method and theoretical basis for promoting the healthy development of open innovation communities,and has guiding significance for companies to optimize community management,quickly screen for ideas,and increase user stickiness.