This Project is on Anomaly detection in multiple time-series. Given more than one time-series with dependency on each other we try to find anomaly in time-series.
This project is developed by Kapil Thakkar and Reshma Kumari, IIT Delhi. (CSE M. Tech. 2014-2016), under the guidance of Dr. Aaditeshwar Seth
-
To refer our work, see our Thesis here:
- thesis/Thesis.pdf
-
To refer code/library refer folder:
- library/
-
To see detailed documentation of library functions refer document:
- library_documentation/Anomaly Detection.pdf
-
To refer data used by us, refer folder:
- database_backup/
-
To see our results refer folder:
- analysis/ Documentation can be found here:
- analysis/resultDoc/results.pdf
-
Work done to collect news articles, fetching links from html search pages, fetching text and related data using AlchemyAPI and Diffbot can be found here:
- newsArticleWork
-
For graph based anomaly, R packages are present in folder:
- mtp/library/R_Packages/
-
Matlab Script for mapping mandis to center using voronoi diagram can be found here:
- mtp/matlabScript/clustering.m
-
Crawlers written to fetch onion data are present here:
- Wholesale Price and Arrival: mtp/oniondataCollection/WholeSalePriceCrawler
- Retail Price: mtp/oniondataCollection/RetailPriceCrawer
- Location of Mandis and Centers: mtp/oniondataCollection/LocationScript
- Required JAR files: mtp/oniondataCollection/jarFiles