In the context of information overload,companies often struggle to effectively identify valuable ideas on their open innovation platforms.In this article,we propose an idea adoption strategy based on machine learning....In the context of information overload,companies often struggle to effectively identify valuable ideas on their open innovation platforms.In this article,we propose an idea adoption strategy based on machine learning.We used data from a well-known open innovation platform,Salesforce,and extracted characteristic variables using the Information Adoption Model.Four classification models were then constructed based on AdaBoost,Random Forest,SVM and Logistic Regression models.Due to significant differences in the number of positive and negative samples in the OIP,we used the SMOTE method to address the problem of data imbalance.The results of the study showed that the ensemble learning models were more accurate in identifying valuable ideas than the individual machine learning models.When comparing the two ensemble learning models,AdaBoost outperformed Random Forest in predicting both positive and negative class samples.The SMOTE-AdaBoost model achieved a recall of 0.93,a precision of 0.92 and an impressive AUC of 0.98 in identifying adopted ideas,which could well identify valuable ideas and has implications for improving the efficiency and quality of idea adoption in OIP.The shortcoming of this work is that it only investigated a single platform.In the future,we will consider extending this method to different platforms and multiple classification problems.展开更多
基金Supported by the National Natural Science Foundation of China(72171090)the Guangdong Basic and Applied Basic Research Fund(2023A1515011551)。
文摘In the context of information overload,companies often struggle to effectively identify valuable ideas on their open innovation platforms.In this article,we propose an idea adoption strategy based on machine learning.We used data from a well-known open innovation platform,Salesforce,and extracted characteristic variables using the Information Adoption Model.Four classification models were then constructed based on AdaBoost,Random Forest,SVM and Logistic Regression models.Due to significant differences in the number of positive and negative samples in the OIP,we used the SMOTE method to address the problem of data imbalance.The results of the study showed that the ensemble learning models were more accurate in identifying valuable ideas than the individual machine learning models.When comparing the two ensemble learning models,AdaBoost outperformed Random Forest in predicting both positive and negative class samples.The SMOTE-AdaBoost model achieved a recall of 0.93,a precision of 0.92 and an impressive AUC of 0.98 in identifying adopted ideas,which could well identify valuable ideas and has implications for improving the efficiency and quality of idea adoption in OIP.The shortcoming of this work is that it only investigated a single platform.In the future,we will consider extending this method to different platforms and multiple classification problems.
