The performance of restricted AIC for irregular histogram models

Area of Science:

• Data Analysis and Visualization
• Statistical Modeling

Background:

• Histograms are essential for preliminary data analysis, including outlier detection and distribution shape identification.
• Selecting an appropriate number and width of bins is critical for accurate histogram interpretation.
• Unequal bin widths can offer advantages over equal widths but present significant challenges in selection.

Purpose of the Study:

• To propose a novel approach for determining optimal bin widths in histograms using the Akaike Information Criterion (AIC).
• To address the difficulties associated with selecting appropriate unequal bin widths for improved data representation.

Main Methods:

• Development of a new AIC-based method tailored for histograms with unequal bin widths.
• Extensive Monte Carlo simulations to evaluate the proposed method's performance.
• Application to empirical datasets to demonstrate practical utility.

Main Results:

• The proposed AIC approach effectively optimizes the selection of unequal bin widths.
• Demonstrated advantages of the novel method over existing techniques through simulations.
• Validation of the method's efficacy using real-world data examples.

Conclusions:

• The novel AIC-based method provides a superior approach for histograms with unequal bin widths.
• This method enhances the reliability of data distribution analysis and outlier identification.
• The approach offers a practical solution to a long-standing challenge in data visualization.