Traffic Inference System Using Correlation Analysis with Various Predicted Big Data

Abstract

Currently, most of the transportation systems require changes to intelligent transportation systems, but most of them focus on efficient transportation rather than on improvement in human life. Sometimes, traffic systems are designed for economic value, and safety-related issues are neglected. A traffic information system that reflects various kinds of environmental information related to people’s safety must be able to reflect not only the existing economic goals but also a safe traffic environment. The traffic environment can be thought of as safety and direct information such as rainfall, including information on specific days when many people are scheduled to be gathered for certain events nearby. Intelligent transportation systems using this information can provide safety-related information for traveling to a specific area or for business trips. In addition, traffic congestion is a social problem and is directly related to a comfort life for individuals. Therefore, addressing various social and environmental factors could make human life more stable and reduce stress as a result. To do that, we need to estimate the impact on traffic based on environmental Big Data. The data can generally be divided into structured data and unstructured data. In inference, structured data analysis is relatively easy due to the precise meaning of the data. Nonetheless, it can be very difficult to predict environmentally sensitive data, such as traffic volume in intelligent transportation systems. To cope with this problem, there are a few systems for handling unstructured data to find out specific events that affect the traffic volume and improve its reliability. This paper shows that it is possible to estimate the exact volume of traffic using correlation analysis with various predicted data. Thus, we may apply this technique to the existing intelligent transportation system to predict the exact volume of traffic with environmentally sensitive data including various unstructured data.