To use a custom kernel in LIBSVM with -t 4 (precomputed kernel), you must first construct an n×(n+1)n \times (n+1)n×(n+1) kernel matrix for training, where the first column contains the sample indices and the rest is the kernel values K(xi,xj)K(x_i, x_j)K(xi​,xj​). For prediction, generate an m×(n+1)m \times (n+1)m×(n+1) matrix with test-to-train kernel values. Ensure the kernel is symmetric and positive semi-definite. We used similar matrix transformations in our Deep Q-Learning with optimization-driven detection models for efficient anomaly classification.

My Article in Deep Q-Learning

Article Deep Guard: Fortifying Digital Authenticity with Deep Q-Lear...