Analysis of Innovative Product Experience Based on Logistic Regression with Data Augmentation

Yuer Yang, Shanlin Zhu, Liangtian Zhao, Yehua Huang, Zhiying Cui, Ying Hong, Xiangjie Wang

Abstract: To investigate the preference of the young for an innovative bank applet named "Cute Cat" and developed by Webank, we design a set of questionnaires to conduct the survey aimed at aesthetics, innovation and practicability. The questionnaire also collected information on students’ age, occupation, and income range to facilitate further analysis. After ten days of data collection, cleaning, and augmentation, we obtained 640 valid data. We conducted the analysis using a logistic regression model, and the results were obtained after adjusting the relevant parameters. We used the median as the threshold for positive reviews. The experimental results show that whether or not the positive reviews are given has a great statistical correlation with the aesthetics, innovation, practicality of the product and the age group and gender of students (H 2 < 0.001). In contrast, the overall impression of the product had almost no correlation with factors such as the income range and monthly living expenses of students (P > 0.05). In conclusion, we summarize or propose four types of data augmentation methods and select the most suitable one to perform Logistic Regression. Results show that the innovation of the product plays a decisive role in the overall experience. In general, the product is favored by college and middle school students. Therefore, the product’s longevity depends on whether its practicality and aesthetics can keep up with the aesthetic trend of this user group.