Main Trend Extraction Based on Irregular Sampling Estimation and Its Application in Storage Volume of Internet Data...

Area of Science:

• Computer Science
• Data Science
• Time Series Analysis

Background:

• Internet data center storage volume is a critical time series for business value.
• Real-world storage data is often "dirty," containing noise, missing values, and outliers, hindering accurate prediction.
• Extracting the main trend from such data is essential for reliable future forecasting.

Purpose of the Study:

• To develop and validate an irregular sampling estimation method for extracting the main trend of noisy time series data.
• To improve the accuracy of future storage volume predictions for internet data centers.

Main Methods:

• Utilized a Kalman filter to effectively remove noise, missing data, and outliers from the time series.
• Employed cubic spline interpolation and an averaging method to reconstruct the primary trend of the storage volume data.
• Applied the developed method to real-world internet data center storage volume data.

Main Results:

• The proposed method accurately estimates the main trend of internet data center storage volume time series.
• Experimental results demonstrate the effectiveness of the developed technique in handling "dirty" data.
• The method significantly contributes to more accurate future storage volume predictions.

Conclusions:

• The irregular sampling estimation method is a robust approach for trend extraction in noisy time series.
• Accurate trend extraction is crucial for reliable forecasting of internet data center storage volumes.
• This research offers a valuable contribution to time series analysis and data center management.