def main(): # X, types, y = ToFormNumpy("D:\\german.txt") X, types, y = ToFormNumpy("D:\\german1.txt") # X, types, y = ToFormNumpy("D:\\tanlanmalar\\german.txt") #minmax_scale(X, copy=False) #Normalizing_Estmation(X, y) y -= 1 _, ln = np.unique(y, return_counts=True) w = Lagranj1(X, y) print(w) res = compactness(X, y, types=types, metric=1) print(res) while X.shape[1] > 2: cond = w != w.min() X = X[:, cond] w = w[cond] res = compactness(X, y, types=types, metric=1) print(res)
def main(): #X, types, y = ToFormNumpy("D:\\tanlanmalar\\gasterlogy1394.txt") #X, types, y = ToFormNumpy("D:\\tanlanmalar\\spame.txt") #X, types, y = ToFormNumpy("D:\\tanlanmalar\\MATBIO_MY.txt") X, types, y = ToFormNumpy( r"D:\Nuu\AI\Selections\Amazon_initial_50_30_10000\data.txt") metric = 1 minmax_scale(X, copy=False) w = Lagranj_nd(X, y) value = w.min() X_Test = np.array(X[:, w == value]) types_Test = np.array(types[w == value]) i = 0 while X_Test.shape[1] < 2000: value = np.min(w[w > value]) X_Test = X[:, w <= value] types_Test = types[w <= value] noisy = find_noisy(X_Test, y, types=types_Test, metric=metric) cond = np.logical_not(noisy) print("\nnoisy = ", len(noisy[noisy == True])) compactness(X_Test[cond], y[cond], types=types_Test, metric=metric) i += 1
def main(): path_train = r"D:\Nuu\AI\Selections\gene-expression\data_train.csv" path_test = r"D:\Nuu\AI\Selections\gene-expression\data_test.csv" X_train, types_train, y_train = ReadFromCSVWithHeaderClass(path_train) X_test, types_test, y_test = ReadFromCSVWithHeaderClass(path_test) #minmax_scale(X_train, copy=False) #minmax_scale(X_test, copy=False) Normalizing_Estmation(X_train, y_train, types_train) print(X_train) _, ln = np.unique(y_train, return_counts=True) w = Lagranj1(X_train, y_train) value = w.max() cond = w == value while len(cond[cond == True]) <= X_train.shape[1]: value = np.max(w[w < value]) cond = w >= value compactness(X_train[:, cond], y_train) return 0 #cond = [356, 2266, 2358, 2641, 4049, 6280] #cond = [356, 2266, 2358, 2641, 2724, 4049] #cond = [356, 2266, 2641, 3772, 4049, 4261] #cond = [4847] #X_train = X_train[:, cond] #X_test = X_test[:, cond] nnc1 = NearestNeighborClassifier_(noisy=True) nnc2 = NearestNeighborClassifier() nnc3 = TemplateClassifier(noisy=True) nn = MLPClassifier() svm = SVC() nnc1.fit(X_train, y_train) nnc2.fit(X_train, y_train) nnc3.fit(X_train, y_train) svm.fit(X_train, y_train) nn.fit(X_train, y_train) mean1 = nnc1.score(X_test, y_test) mean2 = nnc2.score(X_test, y_test) mean3 = nnc3.score(X_test, y_test) mean4 = svm.score(X_test, y_test) mean5 = nn.score(X_test, y_test) print("NearestNeighborClassifier_", mean1) print("NearestNeighborClassifier", mean2) print("TemplateClassifier", mean3) print("SVC", mean4) print("MLPClassifier", mean5)
def main(): #X, types, y = ToFormNumpy("D:\\tanlanmalar\\gasterlogy1394.txt") X, types, y = ToFormNumpy(r"D:\Nuu\AI\Selections\Amazon_initial_50_30_10000\data.txt") #y[y == 2] = 1 minmax_scale(X, copy=False) #minmax_scale(X, copy=False) w = Lagranj_nd(X, y) #return 0 value = w.max() cond = w == value while len(cond[cond == True]) < 5000: value = np.max(w[w < value]) cond = w >= value if len(cond[cond == True]) > 154: compactness(X[:, cond], y, types) return 0 print(len(cond[cond == True])) X = X[:, cond] types = types[cond] metric = 1 #nnc = NearestNeighborClassifier_(noisy=True) nnc = NearestNeighborClassifier() # nnc = TemplateClassifier(noisy=True) nn = MLPClassifier() svm = SVC() k = 10 mean1 = 0 mean2 = 0 mean3 = 0 for i in range(k): X_train, X_test, y_train, y_test = train_test_split( X, y, test_size=0.5, random_state=None, shuffle=True) nnc.fit(X_train, y_train) svm.fit(X_train, y_train) nn.fit(X_train, y_train) mean1 += nnc.score(X_test, y_test) mean2 += svm.score(X_test, y_test) mean3 += nn.score(X_test, y_test) mean1 /= k mean2 /= k mean3 /= k print(mean1, mean2, mean3)
def main(): X, types, y = ToFormNumpy("D:\\tanlanmalar\\IT_BORI_42_6.txt") #X, types, y = ToFormNumpy("D:\\tanlanmalar\\giper_my.txt") y -= 1 minmax_scale(X, copy=False) #Normalizing_Estmation(X, y) compactness(X, y, types=types, metric=1)
def calculate(self): try: noisy_indx = find_noisy(self.X, self.y) noisies = self.count(noisy_indx, self.y) if self.cal > 0: X = self.X[noisy_indx == False] y = self.y[noisy_indx == False] else: X = self.X y = self.y shells = self.count(find_shell(X, y), y) etalons = self.count(find_standard(X, y), y) comp = compactness(X, y) self.result.setRowCount(self.cal + 1) self.result.setItem(self.cal, 0, QTableWidgetItem(str(shells[0]))) self.result.setItem(self.cal, 1, QTableWidgetItem(str(shells[1]))) self.result.setItem(self.cal, 2, QTableWidgetItem(str(noisies[0]))) self.result.setItem(self.cal, 3, QTableWidgetItem(str(noisies[1]))) self.result.setItem(self.cal, 4, QTableWidgetItem(str(etalons[0]))) self.result.setItem(self.cal, 5, QTableWidgetItem(str(etalons[1]))) self.result.setItem(self.cal, 6, QTableWidgetItem(str(comp[0]))) self.result.setItem(self.cal, 7, QTableWidgetItem(str(comp[1]))) self.result.setItem(self.cal, 8, QTableWidgetItem(str(comp[2]))) self.cal += 1 except Exception as exc: QMessageBox.about(self, "Hisoblashda xatolik bor: ", str(exc))
def main(): X, types, y = ToFormNumpy("D:\\tanlanmalar\\gasterlogy1394.txt") X, types, y = ToFormNumpy("D:\\tanlanmalar\\spame.txt") # X, types, y = ToFormNumpy("D:\\tanlanmalar\\MATBIO_MY.txt") minmax_scale(X, copy=False) res = compactness(X, y, types=types, metric=1) print(res[0], res[1], res[2])
def main(): path = r"D:\Nuu\AI\Selections\Amazon_initial_50_30_10000\data.txt" X, types, y = ToFormNumpy(path) minmax_scale(X, copy=False) _, ln = np.unique(y, return_counts=True) w = Lagranj1(X, y) compactness(X, y, types=types, metric=1) while X.shape[1] > 2: cond = w != w.min() X = X[:, cond] w = w[cond] compactness(X, y, types=types, metric=1)
def main(): path = r"D:\Nuu\AI\Selections\LSVT_voice_rehabilitation\data.txt" X, types, y = ToFormNumpy(path) #minmax_scale(X, copy=False) Normalizing_Estmation(X, y) _, ln = np.unique(y, return_counts=True) w = Lagranj1(X, y) compactness(X, y, types=types, metric=1) while X.shape[1] > 2: cond = w != w.min() X = X[:, cond] w = w[cond] compactness(X, y, types=types, metric=1)
def main(): path = r"D:\Nuu\AI\Selections\gene-expression\data1.csv" X, types, y = ReadFromCSVWithHeaderClass(path) #minmax_scale(X, copy=False) Normalizing_Estmation(X, y) _, ln = np.unique(y, return_counts=True) w = Lagranj1(X, y) compactness(X, y, types=types, metric=1) while X.shape[1] > 2: cond = w != w.min() X = X[:, cond] w = w[cond] compactness(X, y, types=types, metric=1)
def main(): path = r"D:\tanlanmalar\GIPER_MY.txt" X, types, y = ToFormNumpy(path) y -= 1 minmax_scale(X, copy=False) # Normalizing_Estmation(X, y) print(compactness(X, y, types)) res = find_standard(X, y, types) res = find_noisy(X, y, types) s = 0 for i in range(res.shape[0]): if res[i] == True and y[i] == 1: print(i + 1) s += 1 print(s) return 0 #nnc = NearestNeighborClassifier_(noisy=True) #nnc = NearestNeighborClassifier() #nnc = TemplateClassifier(noisy=True) nn = MLPClassifier() svm = SVC() k = 10 mean1 = 0 mean2 = 0 mean3 = 0 for i in range(k): X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=None, shuffle=True) nnc.fit(X_train, y_train) svm.fit(X_train, y_train) nn.fit(X_train, y_train) mean1 += nnc.score(X_test, y_test) mean2 += svm.score(X_test, y_test) mean3 += nn.score(X_test, y_test) mean1 /= k mean2 /= k mean3 /= k print(mean1, mean2, mean3)
def main(): path = r"D:\Nuu\AI\Selections\leukemia\leukemia_big.csv" X, types, y = ReadFromCSVWithHeaderClass(path) #minmax_scale(X, copy=False) Normalizing_Estmation(X, y) _, ln = np.unique(y, return_counts=True) w = Lagranj1(X, y) while X.shape[1] > 13: cond = w != w.min() X = X[:, cond] w = w[cond] #print(X.shape) compactness(X, y, types=types, metric=1) """nnc1 = NearestNeighborClassifier_(noisy=True)
def main(): #X, types, y = ToFormNumpy("D:\\tanlanmalar\\gasterlogy1394.txt") #X, types, y = ToFormNumpy("D:\\tanlanmalar\\spame.txt") #X, types, y = ToFormNumpy("D:\\tanlanmalar\\MATBIO_MY.txt") X, types, y = ToFormNumpy("D:\\tanlanmalar\\arcene_train.txt") minmax_scale(X, copy=False) w = Lagranj_nd(X, y) print(w.shape) X_Test = np.array(X[:, w == w.min()]) types_Test = np.array(types[w == w.min()]) print(X_Test) res = compactness(X_Test, y, types=types_Test, metric=1)
def main(): #X, types, y = ToFormNumpy("D:\\tanlanmalar\\gasterlogy1394.txt") X, types, y = ToFormNumpy( r"D:\Nuu\AI\Selections\Amazon_initial_50_30_10000\data.txt") #y[y == 2] = 1 minmax_scale(X, copy=False) #minmax_scale(X, copy=False) w = Lagranj_nd(X, y) unique, ln = np.unique(y, return_counts=True) number_class = len(unique) #return 0 value = w.max() cond_opt = w == value comp_opt = compactness(X[:, cond_opt], y, types) print(len(cond_opt[cond_opt == True]), comp_opt, sep="\t") while len(cond_opt[cond_opt == True]) < X.shape[1]: value = np.max(w[w < value]) cond_current = np.logical_or(w == value, cond_opt) comp_current = compactness(X[:, cond_current], y, types) if comp_opt[number_class] < comp_current[number_class]: cond_opt = cond_current comp_opt = comp_current print(len(cond_opt[cond_opt == True]), comp_opt, sep="\t") return 0 X = X[:, cond] types = types[cond] metric = 1 #nnc = NearestNeighborClassifier_(noisy=True) nnc = NearestNeighborClassifier() # nnc = TemplateClassifier(noisy=True) nn = MLPClassifier() svm = SVC() k = 10 mean1 = 0 mean2 = 0 mean3 = 0 for i in range(k): X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.5, random_state=None, shuffle=True) nnc.fit(X_train, y_train) svm.fit(X_train, y_train) nn.fit(X_train, y_train) mean1 += nnc.score(X_test, y_test) mean2 += svm.score(X_test, y_test) mean3 += nn.score(X_test, y_test) mean1 /= k mean2 /= k mean3 /= k print(mean1, mean2, mean3)