Open AccessOutlier Detection in Climatology Time Series with Sliding Window Prediction
Author(s) -
Manish Mahajan,
Santosh Kumar,
Bhasker Pant
Publication year - 2019
Publication title -
international journal of innovative technology and exploring engineering
Language(s) - English
Resource type - Journals
ISSN - 2278-3075
DOI - 10.35940/ijitee.j9123.0881019
Subject(s) - outlier , sliding window protocol , series (stratigraphy) , time series , computer science , anomaly detection , data mining , window (computing) , value (mathematics) , artificial intelligence , algorithm , machine learning , paleontology , biology , operating system
It is important to identify outliers for climatology series data. With better quality of data decision capability will improve which in turn will improve the complete operation. An algorithm utilising the sliding window prediction method is being proposed to improve the data decision capability in this paper. The time series are parted in accordance with the size of sliding window. Thereafter a prediction model is rooted with the help of historical data to forecast the new values. There is a pre decided threshold value which will be compared to the difference of predicted and measured value. If the difference is greater than a predefined threshold then the specific point will be treated as an outlier. Results from experiment are showing that the algorithm is identifying the outliers in climatology time series data and also remodeling the correction efficiency.