• This repository contains code for two models(LSTM and DBSCAN). I did this in order to find fraud data in the collected data which come from air monitoring station.

• I read many papers in air pollution audit area and i did not find enough papers to do the analysis of fraud data, they just described why they chose one model and made evaluations for the model they chose. Therefore, i made some adjustments to the code in order to get the results of specific fraud data that we detected by our models. In this way, we can help auditors to narrow down auditing contents and save much time.

(I am still working on these models to improve the detecting accuracy, will upload the completed version later...)

• Overview I used LSTM and DBSCAN to do anomaly detection for time series air pollution data.

• Quickstart Python3

• Data I apply my method on the pollution datasets.