The Use of Predictive Modeling in the Evaluation of Technical Acquisition Performance Using Survival Analysis

Journal of Data Science

article

David E. Booth ; Ceyhun Ozgur

关键词: Innovation Performance;    Acquisitions;    Innovation Knowledge;    Hazards Model;    Cox Proportional Hazards;   

DOI  :  10.6339/JDS.201907_17(3).0003

学科分类：土木及结构工程学

来源: JDS

【 摘 要 】

In the recent statistical literature, the difference between explanatory and predictive statistical models has been emphasized. One of the tenets of this dichotomy is that variable selection methods should be applied only to predictive models. In this paper, we compare the effectiveness of the acquisition strategies implemented by Google and Yahoo for the management of innovations. We argue that this is a predictive situation and thus apply lasso variable selection to a Cox regression model in order to compare the Google and Yahoo results. We show that the predictive approach yields different results than an explanatory approach and thus refutes the conventional wisdom that Google was always superior to Yahoo during the period under consideration.

【 授权许可】

CC BY