Classification of functional data: A segmentation approach
Computational Statistics & Data Analysis
Functional Pattern Recognition of 3D Laser Scanned Images of Wood-Pulp Chips
IbPRIA '07 Proceedings of the 3rd Iberian conference on Pattern Recognition and Image Analysis, Part I
Analysis of Time Series with Artificial Neural Networks
MICAI '08 Proceedings of the 2008 Seventh Mexican International Conference on Artificial Intelligence
Functional support vector machines and generalized linear models for glacier geomorphology analysis
International Journal of Computer Mathematics - RECENT ADVANCES IN COMPUTATIONAL AND APPLIED MATHEMATICS IN SCIENCE AND ENGINEERING
Dual features functional support vector machines for fault detection of rechargeable batteries
IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews - Special issue on information reuse and integration
Hi-index | 0.98 |
A statistical analysis of functional data, obtained as reflectance values measured using a hyperspectral sensor, was used to determine water content in vine leaves. Our study was conducted using a sample of 80 vine leaves whose water content was determined by calculating the weight difference between leaves before and after drying in an oven. Two regression models, one linear and the other non-linear, were evaluated and compared: functional linear regression and regression with functional radial basis functions. Compared to traditional methods based on calculating indices that only consider reflectance values in specific wavelengths, the functional approach enables a specific bandwidth or the entire electromagnetic spectrum recorded by the sensor to be taken into account; in other words, the functional approach enables the spectral signature of the leaves to be used. The optimal parameters for each model were determined using a cross-validation procedure and the validity of the approach was tested on a test set drawn from the initial sample. The results obtained demonstrate that water content can be predicted for vine leaves on the basis of their spectral signature, allowing for a certain margin of error. This error was smaller for the non-linear model compared to the linear model.