Idea's Adoption Prediction On Open Innovation Platform

The rapid upgrading of science and technology and the rapid flow of resources have brought convenience to enterprises,but also challenged their innovation capabilities.Open innovation,which advocates open and win-win,multilateral joint for innovation,is sought after by enterprises.Open Innovation Platform(OIP)is a practice for enterprises to implement open innovation.More and more innovative users put forward their own ideas on the platform and actively interact on the platform.For the managers of the platform,active innovation users and effective creativity help to promote the innovation of enterprises.In the face of increasing users and ideas,how to effectively manage users and ideas,accelerate the input of innovation sources into enterprises,and improve management efficiency is especially important.This article selects the open innovation platform salesforce as the research object,aids to building creative adoption prediction models of salesforce.Firstly,qualitative and quantitative methods are used to comprehensively identify the value of users,and subdivide them with a two-dimensional matrix of contribution.From the perspective of innovation contribution,drawing on the RFM value model,the user's innovation proximity,value,and acceptance indicators are selected to calculate the weighted innovation contribution of each.From the perspective of user interaction contribution,interaction depth and breadth are considered to select user interaction behavior indicators,and K-Means clustering algorithm is used to divide categories based on user interaction contribution.On this basis,a two-dimensional matrix of user contributions based on innovation contribution and interaction contribution was constructed,and the platform's users were divided into four categories,namely core users,effective innovation users,active social users,and edge users,providing a theoretical approach for enterprises to identify user value and understand the platform user structure.Secondly,around the creative adoption management of OIP and Triadic Reciprocal Determinism,this article uses the random forest to filter features from the perspective of the interaction between the individual user,platform environment and creative behavior,and builds an creative adoption prediction model based on Adaboost algorithm;Finally,it was compared with the prediction results of traditional logistic regression LR,support vector machine SVM,and BP neural network algorithms,and the validity of the model is verified based on user segmentation.The results found that:(1)There is a large imbalance in the user structure of users.Among the four categories of core users,effective innovative users,active social users,and edge users,active social users account for only 3.48%,while the vast majority are edge users,accounting for 76.53%;(2)Compared with other prediction models,the Adaboost model has the largest AUC value in the imbalanced data sample of this paper,which is 0.83,which shows better accuracy and can more accurately predict creative adoption;(3)Based on user segmentation,the prediction model further enhances the effect.After segmentation,each indicator has a significant improvement compared to the overall forecast.As far as the AUC area,the AUC area among effective innovative users is the highest,reaching 0.98,which is an increase of 18 %;In terms of accuracy,effective innovation users and edge users reach 0.99,which shows that the Adaboost model based on segmented users has better performance.