Parallel recommendation method integrating emotion calculation and network crowdsourcing

The invention discloses a parallel recommendation method integrating emotion calculation and network crowdsourcing. The method generally comprises steps that a user intention detection method based onexpression recognition is accurate and efficient; a collaborative recommendation algorithm supporting user intention detection and network crowdsourcing is supported; the invention discloses a speculation parallelization method of a recommendation algorithm on a distributed platform. The method is based on user expression recognition, deep learning, speculation multithreading technology and mathematical modeling , the expression of a user and the social network characteristics are explored and utilized to construct an interest model of a target user; the recommendation method conforming to the dynamic interest of the user on the distributed platform is studied, the real-time, efficient and parallel recommendation method is constructed, execution efficiency of the recommendation system isimproved, and necessary an