ACCIDENT RISK DIAGNOSIS METHOD, ACCIDENT RISK DIAGNOSIS DEVICE, AND ACCIDENT RISK DIAGNOSIS SYSTEM

The purpose of the invention is to achieve a framework for accurately estimating the accident risk of a driver. This accident risk diagnosis device: stores travel data, which is data indicating a travel state of each of a plurality of vehicles for each time of day, and information pertaining to past accidents of the plurality of vehicles; extracts, from the travel data, the travel data for a prescribed period prior to the day of the accident and travel data for the day of the accident; generates, on the basis of the extracted travel data, training data for each trip, which is a period from starting of a vehicle until a completion; uses the training data to generate an accident risk diagnosis model, which is a machine learning model for diagnosing accident risk; diagnoses an accident risk by inputting to the accident risk diagnosis model a feature amount of a trip of a vehicle for which the accident risk is to be diagnosed; and transmits the diagnosis result to an onboard device installed in the vehicle. 本发明实现