get_Q() Sample Clauses

get_Q(). This function produces a vector with length size equal to the number of data instances. Note that within a loop of the optimization solver we compute the formula presented in the description of the kernel computation. During this loop index iremains constant and j receives all values from 0 up to the number of data instances. Thus, the kernel computation is performed as many times as the number of rows (data instances) of the input file. However, in the get_Q function we also add the labels of the respective vectors so as to compute the following formula: In the above formula K(xi,xj) is equal to the expression in the description of the kernel computation and yi, yj are the labels of the respective vectors. Note that y may either equal to 1 or -1, when we are performing binomial classification, or it may receive other values, when we are performing multi-class classification. We studied binomial classification, in accordance with WP2. Furthermore, due to the existence of the exponent in K(xi,xj) the expression presented in get Q cannot be entirely implemented on hardware.