EMERGENCY DEMAND PREDICTION DEVICE, EMERGENCY DEMAND PREDICTION METHOD AND PROGRAM

To provide an emergency demand prediction device, an emergency demand prediction method and a program which efficiently predict the number of emergency requests which will occur within an object area.SOLUTION: An emergency demand prediction device 1 predicts the number of emergency requests which will occur within an object area. The emergency demand prediction device 1 acquires record data including information of dates when emergency requests occurred, information of locations where the emergency requests occurred, and information of diseases being causes of the emergency requests, generates a first learning model which receives first learning data generated on the basis of record data for learning to output a disease group, and generates a second learning model which receives second learning data generated on the basis of record data for learning and the disease group outputted from the first learning model to output the number of emergency requests which will occur in each of unit areas, and inputs data for prediction which is generated on the basis of record data for prediction and the disease group outputted from the learned first learning model, to the learned second learning model to predict the number of emergency requests which will occur in each of the unit areas.SELECTED DRAWING: Figure 1 【課題】対象エリア内の救急要請の発生数を効率的に予測する救急需要予測装置、救急需要予測方法及びプログラムを提供する。【解決手段】救急需要予測装置１は、対象エリア内の救急要請の発生数を予測する。救急需要予測装置１は、救急要請が発生した日時の情報と、救急要請が発生した位置の情報と、原因となった傷病の情報とを含む実績データを取得し、学習用の実績データをもとに生成される第１の学習用データを入力として傷病群を出力する第１の学習モデルを生成し、学習用の実績データと第１の学習モデルから出力される傷病群とをもとに生成される第２の学習用データを入力として、単位領域ごとの救急要請の発生数を出力する第２の学習モデルを生成し、予測用の実績データと学習済みの第１の学習モデルから出力される傷病群とをもとに生成される予測用データを、学習済みの第２の学習モデルに入力して、単位領域ごとの救急要請の発生数を予測する。【選択図】図１