From: Celebrity profiling through linguistic analysis of digital social networks
Logistic regression | Multinomial naive Bayes | ||
---|---|---|---|
Parameter | Value | Parameter | Value |
Penalty to minimize cost function (penalty) | l2 | Additive smoothing (alpha) | Laplace |
Dual formulation (dual) | Primal formulation | Learn class prior probabilities (fit_prior) | True |
Tolerance level for stopping criteria (tol) | 1.00E-04 | Â | Â |
Relative strength of regularization (c) | 1 | Â | Â |
Calculate the intercept (fit_intercept) | True | Â | Â |
Intercept scaling (intercept_scaling) | 1 | Â | Â |
Pseudo-random number generator (random_state) | 0 | Â | Â |
Solver algorithm (solver) | L-BFGS | Â | Â |
Number of maximum iteration(max_iter) | 100 | Â | Â |
Approach for handling multiple classes (multi_class) | multinomial | Â | Â |