import numpy from sklearn.model_selection import train_test_split from sklearn.linear_model import LinearRegression import time from random import randint import analise_auxiliar array_passe: numpy.ndarray = numpy.concatenate([ analise_auxiliar.get_array_from_pattern( "ROBOCUP-2021-VIRTUAL/DIVISION-B/ER_FORCE/ATA/*Pass.csv"), analise_auxiliar.get_array_from_pattern( "ROBOCUP-2021-VIRTUAL/DIVISION-B/KIKS/ATA/*Pass.csv"), analise_auxiliar.get_array_from_pattern( "ROBOCUP-2021-VIRTUAL/DIVISION-B/RoboCin/ATA/*Pass.csv"), analise_auxiliar.get_array_from_pattern( "ROBOCUP-2021-VIRTUAL/DIVISION-B/RoboFEI/ATA/*Pass.csv"), analise_auxiliar.get_array_from_pattern( "ROBOCUP-2021-VIRTUAL/DIVISION-B/TIGERs_Mannheim/ATA/*Pass.csv") ]) X, y = analise_auxiliar.get_x_y_passes(array_passe, 1.12) x_axis: numpy.ndarray = numpy.fromiter(range(0, 500, 1), dtype=numpy.uint16) score_train: numpy.ndarray = numpy.full(x_axis.shape, 0, dtype=numpy.float64) score_test: numpy.ndarray = numpy.full(x_axis.shape, 0, dtype=numpy.float64) cofs = None start: float = time.time() for j, i in enumerate(x_axis):
import numpy from sklearn.model_selection import train_test_split from sklearn.tree import DecisionTreeRegressor from matplotlib import pyplot import joblib import time from random import randint import analise_auxiliar pyplot.style.use('dark_background') array_chute: numpy.ndarray = numpy.concatenate([ analise_auxiliar.get_array_from_pattern( "ROBOCUP-2019/ER_FORCE/ATA/*Chute.csv"), analise_auxiliar.get_array_from_pattern( "ROBOCUP-2019/ZJUNlict/ATA/*Chute.csv") ]) y: numpy.ndarray = array_chute[:, 0] X: numpy.ndarray = array_chute[:, [1, 2, 3]] model_out: DecisionTreeRegressor = DecisionTreeRegressor( criterion='mse', splitter='best', max_depth=3, min_samples_split=100 * 1e-3, min_samples_leaf=100 * 1e-3, min_weight_fraction_leaf=100 * 1e-3, max_features='auto', random_state=randint(0, 1000), max_leaf_nodes=5,
if numpy.size(var) == 0: continue for j, _var in enumerate( distances_weights[i] ): # apply the weight function for each distance if (_var >= MAXIMUM_DISTANCE): distances_weights[i][j] = 0.0001 continue distances_weights[i][j] = 1 - _var / MAXIMUM_DISTANCE return distances_weights array_passe: numpy.ndarray = analise_auxiliar.get_array_from_pattern( "ALL/*Passe.csv") y: numpy.ndarray = array_passe[:, 0] X: numpy.ndarray = array_passe[:, [1, 2, 3, 4, 4, 6, 7, 8]] knn_out: KNeighborsRegressor = KNeighborsRegressor( n_neighbors=20, weights=customized_weights_linear, n_jobs=1).fit(X, y) joblib.dump(knn_out, "models/avaliacao_passe_knn_with_weights.sav") x_axis: numpy.ndarray = numpy.fromiter(range(1, 50), dtype=numpy.uint16) score_train: numpy.ndarray = numpy.full(x_axis.shape, 0, dtype=numpy.float64) score_test: numpy.ndarray = numpy.full(x_axis.shape, 0, dtype=numpy.float64) start: float = time.time()
import numpy from sklearn.model_selection import train_test_split from sklearn.tree import DecisionTreeRegressor import joblib import time from random import randint import analise_auxiliar array_passe: numpy.ndarray = analise_auxiliar.get_array_from_pattern( "LARC-2020-VIRTUAL/ALL/*Passe.csv") X, y = analise_auxiliar.get_x_y_passes(array_passe, 1.02) tree_out: DecisionTreeRegressor = DecisionTreeRegressor( criterion='mse', splitter='best', max_depth=4, min_samples_split=33 * 1e-3, min_samples_leaf=80 * 1e-3, min_weight_fraction_leaf=87e-3, max_features='auto', random_state=38, max_leaf_nodes=6, min_impurity_decrease=0, min_impurity_split=None, presort='deprecated', ccp_alpha=40).fit(X, y) joblib.dump(tree_out, "models/avaliacao_passe_tree.sav") x_axis: numpy.ndarray = numpy.fromiter(range(1, 1000, 10), dtype=numpy.uint16)