Quality Assessment And Classification Of Multivariate Sensor Time Series Data

In the era of big data,data runs through all aspects of experimental research,so it is of great significance to use appropriate methods to evaluate whether data quality meets the needs of subsequent experiments.Data quality assessment provides strong support for data collectors and data users,which is conducive to data users to make reasonable judgments and make effective decisions during the experiment.There are two objectives of data quality assessment.One is to verify whether there is a problem in the hardware equipment through data quality assessment in the data acquisition stage;the other is to effectively screen out the data with poor data quality in the data set,so as to improve the accuracy of the experiment in the subsequent experiment.In this paper,the quality evaluation and classification algorithm of multi-dimensional sensor timing data are studied based on Parkinson data set.Firstly,in order to effectively improve the quality of timing data,a complete data quality evaluation index system was constructed according to the characteristics of sensor timing data,and a weight integration method was proposed.After the evaluation indexes and rules are determined,the score of each evaluation index is calculated by various methods aiming at different attributes,and the weight calculated by relative dominance degree and fuzzy analytic hierarchy process is integrated to construct the weight of each index,which solves the problem that the importance ranking of data indexes is inconsistent due to the different preferences of decision-makers in the decision-making process.Secondly,a new hesitancy fuzzy set is defined,which takes the scores of different indicators as a part of hesitancy fuzzy set,which makes up the defect that the original hesitancy fuzzy set cannot accurately reflect the data problem.Based on the characteristics of the new hesitancy fuzzy set,the deviation of hesitancy fuzzy element is added into the calculation of hesitancy degree,and the calculation method of hesitancy degree is updated.Based on hesitancy degree,the Minkowski distance measure and Minkowski weighted distance measure between hesitancy fuzzy element are defined.Finally,a new method of VIKOR multi-attribute decision making based on new hesitation fuzzy set is proposed to sort the data quality.According to the characteristics of Parkinson’s data set,the three parts of Parkinson’s data set were standardized separately by using the method of separate standardization.The VIKOR model based on the new hesitation fuzzy set was used to calculate the quality score of the data,and the data with low score was screened out.The remaining data was used for classification training.Verify the validity of the data quality assessment method.