Author: Zen and the Art of Computer Programming
1 Introduction
Knowledge Graph (Knowledge Graph) was first proposed by Jure Sifne et al. of Stanford University in 2013. It is a data model for storing, organizing and reasoning complex and changeable network structure information. Its basic feature is that each fact has a unified identifier ID, and various relationships and their edges can be displayed graphically. Therefore, the knowledge graph can not only model entities, but also model the relationship between entities. Knowledge graphs have powerful expressive capabilities, which can be used to represent complex phenomena, and can also help us quickly retrieve, analyze, make decisions, and recommend data. At present, a large number of knowledge graph databases have been developed, including Google Knowledge Graph, Freebase of Stanford Natural Language Processing Laboratory, etc. Among them, the Google knowledge graph is the largest knowledge graph in the world, organized by billions of triples. Its main uses include search, question and answer, recommendation engine, advertising, healthcare, finance and other fields. In addition, the intelligent decision-making system (KGS) based on knowledge graph is also booming. For example, recommend movies, news, or books based on user preferences and needs; provide recommended products based on users' shopping history, location, and preferences; find resonance and emotional tendencies in movie reviews, and provide corresponding suggestions. This article will introduce an intelligent decision-making system based on knowledge graphs. This system mainly uses knowledge graphs to model historical data such as user search behaviors and purchase behaviors. Based on algorithms such as knowledge graph link prediction and path planning, it can reason about unknown information. Data mining, machine learning, and other methods are used to analyze and mine historical data to further determine users' possible ratings and opinions on specific products or services, and give targeted suggestions. To this end, this paper first introduces the relevant background and technical characteristics of knowledge graphs, and then introduces the technical principles and applications of knowledge graphs in four aspects: data modeling, link prediction, path planning, data mining, and machine learning. Finally, combined with actual cases, the combination and application prospects of knowledge graphs and intelligent decision-making systems are expected.