Development of a machıne learnıng-adapted method for prıorıtızıng seısmıc rısk states of reınforced concrete structures