APPARATUS AND METHOD FOR RECOMMENDATION OF CURLING GAME STRATEGY USING DEEP LEARNING

摘要

An apparatus and a method for recommending a strategy of a curling game are provided. The apparatus for recommending a strategy of a curling game comprises: a memory storing a program for analyzing and recommending a curling game applying deep learning and an MCTS (Monte Carlo tree search) technique; and a processor for executing the program stored in the memory. The processor recognizes preset game situation information by recognizing an image photographing an on-going curling game since the program for analyzing and recommending a curling game is executed, builds a strategic recommendation model by learning the game situation information and data collected from record data for a plurality of previously performed curling games through the deep learning and the MCTS technique, and recommends one or more strategies corresponding to the recognized game situation information as predicted information through the strategy recommendation model.