Research And Modeling Of Human Intelligence Of Stroke Order Intuition

Daily human behavior is characterized by high intelligence,which helps individuals solve complex problems.For example,the writing of Chinese characters is the process of covering a single line and connecting multiple line.However,in the process of writing Chinese characters,the time and space are limited,so it is necessary to proceed as quickly and accurately as possible.This is just like a mathematical problem,which is to solve the shortest path that meets specific rules,and the stroke order is the answer to this process.In this paper,we compares Chinese handwriting behavior(the stroke order problem,SO problem)with the approximate mathematical problems(the arc-weighted euclidean shortest path problem,AESP problems)in graph theory.The work of our study includes the following:(1)Collecting data on human in solving SO problems and AESP problems.For the SO problem,the strokes form specific semantic symbols,so we collect data by processing the Chinese character database.For the AESP problem,the line segments follow a random distribution.So we design a gamified interactive environment to collect data.(2)Analyzing the statistical characteristics of human in solving SO problems and AESP problems.We simplify the solving process by proving the Markov property of human in solving process.Based on the data set from step(1),and by the conclusion from the TSP problem and relevant concepts of Chinese characters,we extract global and local characteristics that may affect human in solving the two problems.Then we analyze the tendencies of human in solving SO and AESP problems from different perspectives.(3)Solving the AESP problem by modeling human performance.Based on the data set and relevant conclusions from step(2),we build a model for simulating human performance in solving AESP problems.By comparing the differences between the models and human performance,we discuss whether the strategy of solving SO problem can be applied to the approximate AESP problem.After analyzing human performance in Chinese handwriting behavior and solving approximate mathematical problems,we find that there are different tendencies in different features,but there are also certain commonalities.On the other hand,applying the Chinese handwriting approach to solving approximate mathematical problems results in even better solutions.After comparing and analyzing the model data,the model based on Chinese handwriting behavior can not only better fit human performance,but also produce better solutions than the mathematical problem-based model,with a fit degree(Kendall rank coefficient)8% higher and a superiority degree(PAO value)19% higher.Our study indicate that other mathematical problems need further analysis,as well as exploring the relationship between heuristic strategies and solving strategies of human performance.Research in this field can not only promote the development and optimization of related algorithms,but also expand cognitive research on Chinese characters.