International Journal of Computer Applications
Foundation of Computer Science (FCS), NY, USA
|
Volume 184 - Issue 35 |
Published: Nov 2022 |
Authors: Shruti Devlekar, Venkatesh Rallapalli, Sangeeta Oswal |
![]() |
Shruti Devlekar, Venkatesh Rallapalli, Sangeeta Oswal . Anomaly Detection in Time Series using Unsupervised Machine Learning Approach. International Journal of Computer Applications. 184, 35 (Nov 2022), 7-13. DOI=10.5120/ijca2022922348
@article{ 10.5120/ijca2022922348, author = { Shruti Devlekar,Venkatesh Rallapalli,Sangeeta Oswal }, title = { Anomaly Detection in Time Series using Unsupervised Machine Learning Approach }, journal = { International Journal of Computer Applications }, year = { 2022 }, volume = { 184 }, number = { 35 }, pages = { 7-13 }, doi = { 10.5120/ijca2022922348 }, publisher = { Foundation of Computer Science (FCS), NY, USA } }
%0 Journal Article %D 2022 %A Shruti Devlekar %A Venkatesh Rallapalli %A Sangeeta Oswal %T Anomaly Detection in Time Series using Unsupervised Machine Learning Approach%T %J International Journal of Computer Applications %V 184 %N 35 %P 7-13 %R 10.5120/ijca2022922348 %I Foundation of Computer Science (FCS), NY, USA
Anomaly detection is the process of identifying data points, observations or events which deviate from normal behavior. To detect anomalies, we have used the PyCaret machine learning library. PyCaret is a low code, open-source library that automates our machine learning processes and helps us detect outliers/anomalies Its Anomaly detection and Regression modules contain various machine learning algorithms and frameworks such as XGBoost, CatBoost, Isolation Forest, DBSCAN, etc. It is a deployment-ready library that is easy to use and helps users to perform end-to-end experiments efficiently. In this paper, we applied clustering and regression-based methods on the NAB Twitter dataset for time series anomaly detection. In the regression method, we predicted the tweets, calculated the difference by comparing them with actual tweets, and used the thresholding technique for anomaly detection.