Time-series data prediction problem analysis through multilayered intuitionistic fuzzy sets.

Area of Science:

• Computer Science
• Artificial Intelligence
• Data Science

Background:

• Time-series prediction is crucial for applications like sales forecasting, medical diagnostics, and network management.
• Existing fuzzy logic methods for time-series prediction often neglect indeterminacy or hesitation.
• Telemetry data prediction is vital for network and data center control software.

Purpose of the Study:

• To introduce the concept of intuitionistic fuzzified time series for improved time-series prediction.
• To develop a novel framework for time-series prediction that accounts for non-determinism and hesitation.
• To address the limitations of traditional fuzzy time-series models in handling uncertainty.

Main Methods:

• Proposed an intuitionistic fuzzified time-series prediction framework.
• Utilized intuitionistic fuzzified logical relationships based on time-series data.
• Evaluated the method using two time-sequence datasets and compared it with existing techniques.

Main Results:

• The suggested intuitionistic fuzzified time-series prediction approach demonstrated effectiveness.
• Performance was validated by comparing predicted results against other intuitionistic time-series methods.
• The method showed reduced root-mean-square inaccuracy and averaged prediction errors.

Conclusions:

• Intuitionistic fuzzified time series provide a valuable tool for handling indeterminacy in time-series prediction.
• The proposed framework offers a more robust approach to forecasting, especially in the presence of hesitation.
• The study highlights the significance of incorporating indeterminacy for more accurate time-series predictions in networking and information centers.