Multi-target feature prediction method based on small amount of questionnaire survey data

Abstract

The invention discloses a multi-target feature prediction method based on a small amount of questionnaire survey data. The method comprises the following steps of S1, arranging the questionnaire survey data; s2, performing interpolation on the missing data in the step S1 by using a k-means method; s3, converting the supplemented non-numeric features in the step S2 into numeric features through one-hot coding and a coding label; s4, correlating the numeric features obtained by conversion in the step S3 with any target feature to be predicted to obtain a task; establishing a multi-task supervised learning model based on a randomly configured radial basis network, and predicting a plurality of tasks; s5, obtaining a final prediction result according to an output result of the model established in the step S4; according to the method, the common information existing in the model parameters or the data characteristics is mined by utilizing the relevance among the plurality of target characteristics, so that the problem of insufficient questionnaire survey data is solved.

Description

Technical field [0001] The invention relates to the field of questionnaire data analysis, in particular to a multi-target feature prediction method based on a small amount of questionnaire survey data. Background technique [0002] At present, offline questionnaire surveys are less and less used in real life, and questionnaire surveys are gradually turning from offline to online, mainly for the following reasons: First, the response rate of questionnaires is low, especially for questionnaire surveys There is also the phenomenon of malicious discarding of questionnaires in, which not only wastes the questionnaire, but also causes environmental pollution to a certain extent. Second, the cost of converting the questionnaire content into data is high, including two parts, one is the high printing cost of the questionnaire. The other part is the high labor cost and time cost of questionnaire distribution; third, the amount of data collected in the questionnaire survey is small. The f...

AI Technical Summary

Problems solved by technology

[0002] At present, offline questionnaire surveys are used less and less in real life, and questionnaire surveys are gradually shifting from offline to online, mainly for the following reasons: First, the return rate of questionnaires is low, especially questionnaire surveys There is also the phenomenon of maliciously discarding questionnaires, which not only wastes questionnaires, but also causes environmental pollution to a certain extent; second, the cost of converting questionnaire content into data is high, including two parts, one part is the high cost of printing the questionnaire, The other part is the high labor cost and time cost of questionnaire distribution; third, the amount of data collected by questionnaire survey is small

The first two disadvantages of the above-mentioned offline questionnaire survey will directly lead to the third disadvantage, and the small amount of questionnaire data will directly lead to inaccurate survey results, losing or deviating from the meaning of the questionnaire survey

Images