Research on Depression through Data Mining

Ke Wan Dawei Liu

摘 要：In this paper， we focused on Chinese university students' depression issues and acquired data by a survey in universities. The data illustrates that nearly a half of Chinese university students could face mental health issues. People's living habits can affect their mental states. Students who prefer to have spicy food， adequate sleep and moderate amount of exercise is more likely to have a low prevalence of depression. In particular， spicy food can reduce the incidence of depression to some extent.

關(guān)鍵詞：depression； data mining； university students； prediction； decision tree； clustering.

I. INTRODUCTION

University students belong to one of the major target groups of depression[1]. We used a survey to acquire data in Chinese universities so that we can have a better understanding of the students mental states and some influential factors. The questionnaire includes two sections， one is depression test， the other is living habit collection. We used K-means and C4.5 algorithm to do the analysis.

II.DATA ANALYSIS AND RESULTS

We acquired 350 samples in universities. After data cleaning process， we finally have an effective data set and its size is 341. The Cronbachs Alpha of this data set is 0.874； it means the data set has a high reliability. There are 49.56% respondents facing depression issues and 12% of them are in relatively serious situations. The difference in the incidence of depression at different ages is apparent. People's age is negatively correlated with negative emotions. Age has a great influence on university students self-control， learning ability and adaptability[2].

A.Clustering

After K-means clustering， 4 categories of living habits can be found. Category A： light diet， lack of slee， excessive exercise. Category B： having spicy food， lack of sleeping and exercise. Category C： light diet and lack of sleep and exercise. Category D： spicy diet， adequate sleep and moderate amount of exercise. The mental states of the student with a Category D living habit are better. It is easy to notice that the length of sleep and exercise volume will have an influence on mental health.

B.Decision Tree

We have put the data set into Weka to generate a decision tree. After pruning and interpreting， the final version of the decision tree is shown as figure 1. Testing the tree with the original training data， the accuracy is 70.9%. Testing with 20 new samples， the accuracy is 65%.

The most noteworthy result of this decision tree is the effect of diet on depression.When other conditions are the same， spicy diet samples are not easy to get depression compared with light diet samples. Eating spicy food can promote the release of endorphin and P substance from the brain， so that people can feel pleasure[3]. But there is no exact conclusion of the benefits. So it is not advisable to eat too much spicy food because of the results of this work.

III.CONCLUSION

After all the processes have been done， the following conclusions are obtained. Nearly half of the university students in China may have mental health problems. And the proportions of moderate and severe depression are also quite high. Age will affect the incidence of depression in university students. The younger the student is， the higher is the incidence rate. The living habit with proper spicy food， adequate sleep， and moderate amount of exercise will lead to low depression incidence. Eating spicy food has a moderating effect on emotion， can reduce the incidence of depression.

REFERENCES

[1]Tennant， C.， Life events， stress and depression： a review of recent findings. Australian & New Zealand Journal of Psychiatry， 2015. 36（2）： p. 173-182.

[2]Lammers， W.J.， Academic Success as a Function of the Gender， Class， Age， Study Habits， and Employment of College Students. Research in the Schools， 2001. 8（2）： p. 71-81.

[3]Sutep， G.， Are Rice and Spicy Diet Good for Functional Gastrointestinal Disorders？ Journal of Neurogastroenterology & Motility， 2010. 16（2）： p. 131-138.