CB-TFA to RVM on Large Scale Problems

RVM enables sparse classification and regression functions to be obtained by linearly-weighting a small number of fixed basis functions from a large dictionary of potential candidates. TOA on RVM has O(M3) time and O(M2) space complexity, where M is the training set size. It is thus computationally infeasible on very large data sets. TFA was put forward to overcome this problem ,but it is not perfect to large scale problems. We propose CB-TFA based on TFA. CB-TFA decompose large datasets to data blocks, get the solution by chain iteration, taking TFA as basis algorithm, reduce the time complexity further more while keeping high accuracy and sparsity simultaneously. Regression experiments with synthetical large benchmark data set demonstrates CB-TFA yields state-of-the-art performance.