A comparison between the brand awareness and logistic prediction models for predicting election results

Candidates involved in the elections normally formulate their own strategies through the method of opinion polls, and consequently it is extremely important if a prediction model can effectively compute and provide their supporting rate. A suitable election prediction model should be available and implemented in full practice to satisfy the critical necessity for obtaining a result with accuracy, stability, and simplicity. By using the opinion polls gathered from the elections held in Tainan City, Taiwan, with the assistance of the brand awareness prediction and logistic prediction models frequently adopted in elections, this study tests out the validity of these aforementioned two prediction methods, in terms of the three practical necessities referenced above. The results of this research may imply that the forecasted outcomes obtained from both aforementioned models are very promising in terms of the accuracy and stability. However, the brand awareness prediction model will be better if the simplicity of the calculation process is the key objective.