Independent mood traits comprise three primary components -pleasure, arousal, and dominance (Mehrabian, 1996). Forecasting these traits is beneficial for several subjects, such as behavioral science, cognitive science, decision making, mood disorders treatment, and virtual character development in artificial intelligence. In this study, an extended model is proposed to predict independent mood components based on the emotion and mood history of 108 individuals with different backgrounds and personalities. Emotion history of all these individuals was recorded hourwise for six days, and their daily mood history obtained. The proposed model consists of various types of statistical forecasting methods, such as Holt-Winter's additive model and seasonal time series model, by integrating current known appraisal theories and aided by mood history probability distribution. The predicted values for the seventh day and the trend of the outcome results reveal that: (1) Pleasure mood trait trend varies significantly between individuals, but it can be considered as predictable; (2) Arousal mood trait is unpredictable for a short time interval; however, it is possible to have close predictions over long time intervals. (3) Dominance mood trait can be predicted for a short time interval, but not for a long time interval. These findings can shed light on the way mood states and behavior of human beings can be predicted.As expected, personality traits of these individuals differed significantly from each other. As soon as the OCEAN personality traits of the individuals were obtained, their initial PAD traits were also measured by the mapping explained in section 2.2. For example, for one of the individuals Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism were found to be 0.65, 0.97, 0.35, 0.60, and 0.45, respectively. Therefore, by using the mentioned mapping from OCEAN personality traits to PAD (Mehrabian, 1996; Gebhard, 2004), the participant's initial pleasure, arousal, and dominance traits were found to be 0.51, 0.02, and 0.34, respectively. Table 1 made it possible to create time series data for emotion states regarding PAD mood traits. Different types of time series forecasting models were suggested by using SPSS for each mood component of each individual with 95 percent confidence. It has been observed that the trends for pleasure mood component differed significantly from one person to another, but it had a predictable direction for more than 98 percent of individuals. On
The mapping illustrated in