import data
import helper as h
from sklearn.metrics import classification_report, confusion_matrix, \
accuracy_score, precision_score, f1_score, recall_score
from sklearn.neighbors import KNeighborsClassifier
from sklearn.model_selection import train_test_split
# Adição das classes mediante o volume de vendas
data.classification('../files/input/')
# Criação do data frame a partir do ficheiro csv
df = h.get_data('../files/output/classification.data.csv')
# Criação de data frame com base no data frame train,
# sem a coluna 'Weekly_Sales', visto que é o valor que se pretende prever
X = df.drop(columns=['class', 'Unnamed: 0', 'Dept', 'Type', 'Size'])
# Criação de uma lista com os valores da coluna que se pretende prever
y = df['class'].values.tolist()
# Utilização de um módulo da sklearn para dividir o
# data frame train, em train e test, por forma a avaliar a accuracy
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3)
# Obtenção do melhor valor de K
acc_arr = []
predict_arr = []
conf_matrix = []
class_report = []
prec_score = []
f1 = []
recall = []
for K in range(25):
import numpy as np import neuralNetwork as nn import data #------------------- # 0. ハイパーパラメータの設定 dataType = 5 # データの種類 activeType = 2 # 活性化関数の種類 hDim = 20 # 中間層のノード数 alpha = 1 # 学習率 rate = 0.5 # ノード選択確率（ドロップアウト） #------------------- #------------------- # 1. データの作成 myData = data.classification(negLabel=0, posLabel=1) myData.makeData(dataType=dataType) #------------------- #------------------- # 2. データを学習と評価用に分割 dtrNum = int(len(myData.X) * 0.9) # 学習データ数 # 学習データ（全体の90%） Xtr = myData.X[:dtrNum] Ytr = myData.Y[:dtrNum] # 評価データ（全体の10%） Xte = myData.X[dtrNum:] Yte = myData.Y[dtrNum:] #-------------------
# -*- coding: utf-8 -*- import numpy as np import kernelFunc as kf import kernelSVM as svm import data #------------------- # 1. データの作成 myData = data.classification(negLabel=-1.0, posLabel=1.0) myData.makeData(dataType=5) #------------------- #------------------- # 2. データを学習と評価用に分割 dtrNum = int(len(myData.X) * 0.9) # 学習データ数 # 学習データ（全体の90%） Xtr = myData.X[:dtrNum] Ytr = myData.Y[:dtrNum] # 評価データ（全体の10%） Xte = myData.X[dtrNum:] Yte = myData.Y[dtrNum:] #------------------- #------------------- # 3. 標準化 xMean = np.mean(Xtr, axis=0) xStd = np.std(Xtr, axis=0) Xtr = (Xtr - xMean) / xStd Xte = (Xte - xMean) / xStd #-------------------
import data
import helper as h
from sklearn import tree
from sklearn.preprocessing import Imputer
data.classification()
train = h.get_data('../files/output/classification.data.csv')
test = h.get_data('../files/output/test.csv')
# The columns that we will be making predictions with.
x_columns = [
'Store', 'Dept', 'week_number', 'IsHoliday', 'Type', 'Size', 'Temperature',
'Fuel_Price', 'CPI', 'Unemployment'
]
# The column that we want to predict.
y_column = ["class"]
# Troca os valores nulos que interromperiam o modelo, pela média
# Esta estratégia foi adotada devido ao facto de se tratar das colunas Unemployment e CPI
imp = Imputer(missing_values='NaN', strategy='mean', axis=0)
imp = imp.fit(test[x_columns])
X_test_imp = imp.transform(test[x_columns])
clf = tree.DecisionTreeClassifier()
clf.fit(X=train[x_columns], y=train[y_column].values.ravel())
clf.feature_importances_
predictions = clf.predict(X_test_imp)
test['class'] = predictions