Loading...
Paper Number
2780
Paper Type
Complete
Abstract
We evaluate the value of behavioral data generated by Internet-of-Things (IoT) technologies in demand prediction by focusing on the automotive industry. Specifically, we incorporate driving behavior, i.e., hard braking, hard acceleration, and speeding, into predicting maintenance, repair, and operations (MRO) service demands. We compare a baseline model that includes historical MRO data, driver and vehicle attributes, and GIS data (i.e. weather and traffic density) with our model that includes driving behavior as additional predictive features. Our results show that incorporating driving behavior features improves precision, recall, and F1 score by 3.14%, 10.97%, and 3.91%, respectively. Moreover, all three driving behavior features are among the top six important features in predicting MRO demand. We further show that the predictive power of driving behavior features is greater for predicting maintenance services and old vehicles. Our results shed practical implications on leverage IoT enabled behavior data in MRO prediction.
Recommended Citation
Zhang, Jieyi and Yang, Cenying, "The Value of Internet-of-Things Enabled Behavioral Data in Demand Prediction" (2024). ICIS 2024 Proceedings. 9.
https://aisel.aisnet.org/icis2024/iot_smartcity/iot_smartcity/9
The Value of Internet-of-Things Enabled Behavioral Data in Demand Prediction
We evaluate the value of behavioral data generated by Internet-of-Things (IoT) technologies in demand prediction by focusing on the automotive industry. Specifically, we incorporate driving behavior, i.e., hard braking, hard acceleration, and speeding, into predicting maintenance, repair, and operations (MRO) service demands. We compare a baseline model that includes historical MRO data, driver and vehicle attributes, and GIS data (i.e. weather and traffic density) with our model that includes driving behavior as additional predictive features. Our results show that incorporating driving behavior features improves precision, recall, and F1 score by 3.14%, 10.97%, and 3.91%, respectively. Moreover, all three driving behavior features are among the top six important features in predicting MRO demand. We further show that the predictive power of driving behavior features is greater for predicting maintenance services and old vehicles. Our results shed practical implications on leverage IoT enabled behavior data in MRO prediction.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
Comments
17-IOT