Automatic anomaly detection in warehouse environmental conditions using a hybrid approach based on machine learning

A high percentage of food waste is caused by improper storage conditions. Traditional monitoring systems are often based on manual sampling and asynchronous response, which leads to reduced efficiency. This paper implements an intelligent method for controlling environmental conditions in warehouses, which can help improve product quality and reduce costs by using modern technologies such as machine learning. In the proposed method, temperature, humidity, and pressure prediction are performed using LSTM and anomaly detection is performed by combining Isolation Forest, DBSCAN, and Auto-Encoder. The proposed method increases detection accuracy and reduces false errors.

کلیدواژه ها:

Anomaly detection ، LSTM ، Isolation Forest ، DBSCAN ، Auto-encoder