An insurance company, using artificial intelligence, analyzed a data array including the files of millions of Russian motorists. According to the results of the study, more than 500 different factors were found to be insignificant in determining the risk of getting into an accident.
“Data analysts, having analyzed the insurance history of millions of Russian motorists using ML using several hundred parameters, did not find a correlation between the color of the car, the zodiac sign of its owner or the “dangerous” numbers in the driver’s date of birth according to the numerology matrix with the level of real risks behind the wheel,” notes the Rosgosstrakh release received by Motor.
Data analysis confirmed that accident rates are influenced by completely different factors: the experience and behavior of the driver, the tendency to violate traffic rules, vehicle power, mileage and the specifics of the operating area.
As Olga Veresova, director of the analysis and modeling department of Rosgosstrakh, explained, a modern risk assessment model for calculating a fair tariff takes into account over 100 significant factors. For example, in comprehensive insurance, detailed location plays an important role – the characteristics of the client’s place of residence and regular trips, right down to the analysis of accident rates at specific intersections.
Previously, the 10 most reliable used BMWs were named.
