In the learning process of junior high school students,many educators are aware of the importance of non-intellectual factors.Therefore,whether from a theoretical or experimental point of view,the research on non-intellectual factors is of great significance,and it is also in line with the current mainstream.development trend.For a long time,my country’s education system has focused on cultivating the intellectual development of primary and secondary school students and ignored the positive influence of non-intelligence factors in the learning process.Since the 1980 s,many psychologists in our country have carried out a large number of qualitative and quantitative researches on the concept of "non-intelligence factors",and obtained corresponding research results.The researchers have also achieved valuable research results on the related topics of the influence of non-intelligence factors on the academic performance of middle school students,which provides a theoretical basis for the reform of middle school education and teaching.Rarely.This study takes students from a middle school in Pingdingshan City as the research object.On the basis of existing research results at home and abroad,using methods such as literature research,questionnaire survey,interview survey,etc.Main non-intellectual and intellectual factors such as learning interest,learning self-confidence,learning anxiety,emotional stability,attention stability,competitive spirit,achievement motivation,learning perseverance,learning responsibility,communication motivation,independence,and thirst for knowledge in each grade group and related survey results on the academic performance of middle school students in different grade groups in three subjects.Combined with the survey results and teachers’ suggestions,a variety of machine learning algorithms are used to model the relationship between students’ intelligence factors,non-intelligence factors,and family,and middle school students’ Chinese,math,and English scores.And the non-intelligence factors of different disciplines were analyzed.The specific research results are as follows:(1)Theoretically,the decision tree and random forest with no parameter model should be able to get the best results with few samples,while the parametric models such as multiple linear regression and neural network should be greatly affected,and the prediction results are not good.From the actual research results,the random forest can achieve better results than the decision tree because of its integrated characteristics,while the decision tree will produce deviations and poor results.(2)The performance of different machine learning models is different for different subjects.Random forests are the best in all subjects.Neural network and decision tree are the worst in all subjects.As for multiple linear regression and support vector machine,they show similar predictive ability,which is related to the sample distribution of related subjects.(3)From the research results,different non-intelligence factors have different effects on students’ grades.In the seventh grade,intelligence(reasoning)has the greatest influence,followed by attention stability,learning interest and other factors;In the eighth grade,the influence of intelligence factors(reasoning)weakened,and attention stability began to play the greatest role;In the ninth grade,the influence of intelligence factors(reasoning)further declined.At this time,the most influential factor is competitiveness,and attention stability,learning interest and other factors also have certain influence. |