Latent group detection in functional partially linear regression models.

Area of Science:

• Statistics
• Bioinformatics
• Genetics

Background:

• Functional data analysis is crucial for understanding complex biological processes.
• Heterogeneous relationships between scalar responses and functional covariates require advanced modeling techniques.
• Identifying latent group structures is key to uncovering nuanced biological variations.

Purpose of the Study:

• To propose a functional partially linear regression model with latent group structures.
• To accommodate heterogeneous relationships between scalar responses and functional covariates.
• To identify salinity tolerant barley plants by detecting latent groups.

Main Methods:

• Developed a functional partially linear regression model incorporating latent group structures.
• Employed a K-means clustering-inspired algorithm for latent group identification.
• Established theoretical properties including consistency, convergence rate, and asymptotic distribution.

Main Results:

• The proposed method demonstrated higher accuracy in recovering latent groups compared to existing methods.
• Accurate estimation of functional coefficients was achieved.
• Simulation studies validated the model's performance.

Conclusions:

• The functional partially linear regression model with latent groups effectively handles heterogeneous relationships.
• The method successfully identifies distinct groups within barley families based on salinity tolerance.
• This approach aids in detecting salinity tolerant barley varieties.