Prediction and Monitoring of Ton-Kilometers and Waybill for Detecting Abnormal Behavior
| Autor: | Amirali Eghtesad, Abbas Saghaei, Amir Azizi |
|---|---|
| Jazyk: | perština |
| Rok vydání: | 2023 |
| Předmět: | |
| Zdroj: | چشمانداز مدیریت صنعتی, Vol 13, Iss 4, Pp 249-273 (2023) |
| Druh dokumentu: | article |
| ISSN: | 2251-9874 2645-4165 |
| DOI: | 10.48308/jimp.13.4.249 |
| Popis: | The objective of this research is to predict and monitor the ton-kilometers and waybill on the Iran's roads to find abnormal Behavior. In this study, data was collected from monthly observations over a period of 6 years (1395 to 1400) by the Iran Road Maintenance and Transportation Organization, categorized by province. Different machine learning techniques, deep learning, and time series methods were employed to predict ton-kilometers, and the results were monitored for abnormal behavior following an increase in interest rates and taxes. For model implementation, a dataset of 72 records of ton-kilometers and 72 records of issued waybill, collected from 32 provinces over six years of road transportation, was utilized. Initially, four different prediction methods, including random forest, LSTM neural network, ARIMA, and ETS, were extensively examined. The empirical results indicate that the random forest outperforms the other models. this study employs the statistical quality control tool, the z-score, to detect outliers and abnormal behavior in the data. The empirical findings reveal that out of the 32 provinces, three provinces exhibit abnormal behavior, and one of them is attributed to factors other than an increase in interest rates and transportation taxes. |
| Databáze: | Directory of Open Access Journals |
| Externí odkaz: |
|