Measuring the Accuracy of Prediction in a Simulated Environment

Abstract

Prediction is an important cognitive activity in any decision-making process. For dynamic process control tasks, this activity is crucial, but studying it in a real-life environment is difficult. This article describes how a computerized simulation was used to study prediction in a complex environment. Using the simulation, the effects of many factors on the accuracy of prediction were successfully measured. These factors are mainly the nature of the task the method of presentation of the information, the number of repetitions of a specific task, and the length of the time taken for prediction. Three different tasks and four distinct interfaces were used in the simulation. Prediction was found to be significantly affected by all of the factors except the number of repetitions. Surprisingly, no learning effect was observed throughout the accomplishment of the various tasks.