Application of uncertainty reasoning based on cloud model in time series prediction

Area of Science:

• Computer Science
• Artificial Intelligence
• Data Science

Background:

• Time series prediction is crucial for forecasting in areas like meteorology and finance.
• Traditional methods like exponential smoothing often fail to adapt to changing trends due to fixed weights.

Purpose of the Study:

• To develop an adaptive time series prediction method.
• To improve the accuracy of simple exponential smoothing by incorporating dynamic trend adjustments.

Main Methods:

• Utilized an uncertainty reasoning approach based on the cloud model.
• Implemented a cloud logic controller to dynamically adjust the smoothing coefficient in simple exponential smoothing.
• Applied the method to various time series datasets for validation.

Main Results:

• The proposed cloud model-based method successfully adapted to current time series trends.
• Experimental results demonstrated the validity and improved performance over traditional methods.
• The dynamic adjustment of the smoothing coefficient significantly enhanced prediction accuracy.

Conclusions:

• The cloud model-based uncertainty reasoning offers a robust enhancement for time series prediction.
• This adaptive approach effectively addresses the limitations of fixed-weight methods.
• The technique shows promise for improving forecasting accuracy in diverse applications.