Machine Learning Algorithm Competition Practice

This book is a book that systematically introduces competitions in the field of algorithm competitions. It not only contains the basic theoretical knowledge of competitions, but also elaborates on the ideas and techniques for winning points in competitions based on multiple directions and cases. The book is divided into five parts: the first part focuses on the general process of algorithm competitions, introducing the core content and specific work of each part of the competition; the second part introduces issues related to user portraits; the third part focuses on time series prediction problems, first describing common problem-solving ideas and techniques for such problems, and then analyzing the Global Urban Computing AI Challenge on the Tianchi platform and the Corporación Favorita Grocery Sales on the Kaggle platform Forecasting; the fourth part mainly introduces the core technology and business of computational advertising, including ad recall, ad ranking and ad bidding. Two practical cases are the 2018 Tencent Advertising Algorithm Competition-Similar Population Expansion and the TalkingData AdTracking Fraud Detection Challenge on the Kaggle platform; the fifth part is based on content related to natural language processing. The practical case is the Quora Question Pairs competition on the Kaggle platform. This book is suitable for people who are engaged in algorithmic positions related to machine learning, data mining and artificial intelligence.