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.　©　2024,　Science　China　Press.　All　rights　reserved.