A typical prism-coupled surface-plasmon-resonance biosensor comprises a metal thin film in contact with a solution containing an analyte to be sensed. The metal film also acts as a binding surface for bioreceptor molecules to capture and concentrate the analyte molecules of interest. We investigated the use of a porous, anisotropic, periodically non-homogeneous material called a chiral sculptured thin film (CSTF) grown on top of the metal film to confine the solution in its pores. The efficacy of a basic plasmonic sensor was compared with those of two types of sensors containing a CSTF, one type having a metal-nanoparticle layer at a distance of one period from the metal film and the other without that layer. The chosen analyte was Immunoglobulin G and the chosen bioreceptor was Protein A. Measurements were made over a wide angular range rather than over a small range tied to the excitation of a surface-plasmon-polariton wave, and the collected data were used to train a machine-learning algorithm called a support vector machine for classification. We concluded that the metal/CSTF sensor with a metal-nanoparticle layer performs best, the metal-nanoparticle layer being crucial to its better performance.