Research Article
Comparing Methods of Imputation for Time Series Missing Values
@INPROCEEDINGS{10.1007/978-3-030-94182-6_24, author={Renkang Geng and Mingran Li and Mingxu Sun and Yujie Wang}, title={Comparing Methods of Imputation for Time Series Missing Values}, proceedings={IoT and Big Data Technologies for Health Care. Second EAI International Conference, IoTCare 2021, Virtual Event, October 18-19, 2021, Proceedings, Part II}, proceedings_a={IOTCARE PART 2}, year={2022}, month={6}, keywords={Missing value Filling method Traffic data}, doi={10.1007/978-3-030-94182-6_24} }- Renkang Geng
Mingran Li
Mingxu Sun
Yujie Wang
Year: 2022
Comparing Methods of Imputation for Time Series Missing Values
IOTCARE PART 2
Springer
DOI: 10.1007/978-3-030-94182-6_24
Abstract
Due to the rapid development of modern information engineering, a lot of data are used in machine learning and data cleaning and data mining of the hot research fields, such as a large portion of the data algorithm and related data model are built for complete data set, But in our real life and work, the absence of data exists in a large number of data collection, collation, transmission, storage and other links, it causes many obstacles and difficulties to build a model for complete data. The general way of dealing with missing values for simple delete, that deal with missing value method is a simple convenient but can cause: two aspects of the problem and the inconvenience caused by the original data set to reduce, reduce the reliability of the data, especially in the case of data loss is bigger, can cause a large number of data sets to reduce and missing, This has caused a lot of trouble to our work and research, so we need to find a more efficient and better method than direct deletion. In order to better solve the above problems, we mainly fill in the missing values of time series data, which has become an urgent problem to be solved. In this paper, mean filling, median filling, mode filling, PCA-EM filling and other methods are used to fill traffic data. By comparing these methods, the filling effect of each method is evaluated.
