A new regression model for bimodal data and applications in agriculture

Area of Science:

• Statistics
• Agricultural Science

Background:

• Bimodal data is prevalent in agriculture.
• Existing Gaussian regression models may not adequately capture bimodal distributions.

Purpose of the Study:

• To introduce and evaluate a novel regression model for bimodal data.
• To extend heteroscedastic Gaussian regression for improved applicability in agriculture.

Main Methods:

• Definition of the odd log-logistic exponential Gaussian regression with two systematic components.
• Parameter estimation using the method of maximum likelihood.
• Model assumption validation via case deletion and quantile residuals.

Main Results:

• Simulations indicate accurate maximum-likelihood estimators.
• The proposed model demonstrates suitability for bimodal agricultural data.
• The model's utility is confirmed through real-world agricultural case studies.

Conclusions:

• The odd log-logistic exponential Gaussian regression is a valuable tool for analyzing bimodal agricultural data.
• The method of maximum likelihood provides reliable parameter estimates.
• The model offers a significant advancement for statistical modeling in agriculture.