def Ionosphere(X):
binaryCols = {"signal": {"g": 1, "b": 0}}
X = X.copy()
X = Processor.removeMissing(X)
X = X.drop(columns=['col0', 'col1', "col13"])
X = Processor.toBinaryCol(X, binaryCols)
Y = X["signal"]
X = X.iloc[:, :-1]
return [X, Y]
def ttt(X):
labels = {"o": 0, "b": 1, "x": 2}
encoding = {"result": {"positive": 1, "negative": 0}}
X = X.copy()
X = Processor.toBinaryCol(X, encoding)
X = Processor.OHE(
X, cols=["tl", "tm", "tr", "ml", "mm", "mr", "bl", "bm", "br"])
Y = X["result"]
X = X.drop(columns=["result"])
return [X, Y]
def adult(X):
binaryCols = {
"sex": {
"Male": 0,
"Female": 1
},
"salary": {
">50K": 0,
"<=50K": 1
}
}
X = X.copy()
X = Processor.removeMissing(X)
X = Processor.toBinaryCol(X, binaryCols)
X = Processor.normalize(X, ["fnlwgt", "hours-per-week"])
Y = X["salary"]
X = X.iloc[:, :-1]
X = Processor.OHE(X)
countryCols = [
"native-country_Cambodia", "native-country_England",
"native-country_Puerto-Rico", "native-country_Canada",
"native-country_Outlying-US(Guam-USVI-etc)",
"native-country_India", "native-country_Japan",
"native-country_Greece", "native-country_South",
"native-country_China", "native-country_Cuba",
"native-country_Iran", "native-country_Honduras",
"native-country_Italy", "native-country_Poland",
"native-country_Jamaica", "native-country_Vietnam",
"native-country_Portugal", "native-country_Ireland",
"native-country_France", "native-country_Dominican-Republic",
"native-country_Laos", "native-country_Ecuador",
"native-country_Taiwan", "native-country_Haiti",
"native-country_Columbia", "native-country_Hungary",
"native-country_Guatemala", "native-country_Nicaragua",
"native-country_Scotland", "native-country_Thailand",
"native-country_Yugoslavia", "native-country_El-Salvador",
"native-country_Trinadad&Tobago", "native-country_Peru",
"native-country_Hong", "native-country_Holand-Netherlands"
]
X = X.drop(columns=(["capital-gain", "capital-loss", "education-num"] +
countryCols))
return [X, Y]
def mam(X):
X = X.copy()
X = Processor.fillMissing(X)
Y = X["result"]
X = X.drop(columns=["result"])
return [X, Y]
df = pd.DataFrame(data)
df_to_table(df, 'time_table_all_final')
# print(evaluate_acc(Processor.ToNumpyCol(Y_test), model.predict(X_test.to_numpy())))
#print(cross_validation(5, X_train.to_numpy(), Processor.ToNumpyCol(Y_train), model))
elif ds == "ionosphere":
path = "../datasets/ionosphere/ionosphere.data"
header = ["{}{}".format("col", x) for x in range(33 + 1)]
header.append("signal")
All = Processor.read(path, header)
[X, Y] = Clean.Ionosphere(All)
[X_train, X_test, Y_train, Y_test] = Processor.split(X, Y, train=0.8)
setup = '''
from Project1.src.NaiveBayes import NaiveBayes
from Project1.src.Processor import Processor
from Project1.src.Clean import Clean
from Project1.src.CrossValidation import cross_validation
path = "../datasets/ionosphere/ionosphere.data"
header = ["{}{}".format("col", x) for x in range(33 + 1)]
header.append("signal")
import matplotlib.pyplot as plt
import numpy as np
from Project1.src.LogisticRegression import LogisticRegression
from Project1.src.NaiveBayes import NaiveBayes
from Project1.src.Processor import Processor
from Project1.src.Clean import Clean
from Project1.src.CrossValidation import cross_validation
from Project1.src.CrossValidation import evaluate_acc
print("Analyzing the ionosphere data set")
path = "../datasets/ionosphere/ionosphere.data"
header = ["{}{}".format("col", x) for x in range(33 + 1)]
header.append("signal")
All = Processor.read(path, header)
[X, Y] = Clean.Ionosphere(All)
X = X.to_numpy()
Y = Processor.ToNumpyCol(Y)
iters = np.arange(20, X.shape[0], 50)
#print(X.shape)
#print(Y.shape)
accuracies = []
for iter_ in iters:
#rowsX = X[0:X.shape[0], :]
#rowsY = Y[0:Y.shape[0], :]
rowsX = X[0:iter_, :]
Learning rates and threshold gradient were chosen using the results of the hyperparameter tuning script
"""
from Project1.src.LogisticRegression import LogisticRegression
from Project1.src.NaiveBayes import NaiveBayes
from Project1.src.CrossValidation import cross_validation
from Project1.src.Processor import Processor
from Project1.src.Clean import Clean
from Project1.src.HPTuning import df_to_table
import pandas as pd
# Find accuracies for ionosphere data set
print("Analyzing the ionosphere data set")
path = "../datasets/ionosphere/ionosphere.data"
header = ["{}{}".format("col", x) for x in range(33 + 1)]
header.append("signal")
All = Processor.read(path, header)
[X, Y] = Clean.Ionosphere(All)
ionosphere_results = ['ionosphere']
acc, _, _ = cross_validation(5,
X.to_numpy(),
Processor.ToNumpyCol(Y),
LogisticRegression(),
learning_rate=1.0,
max_gradient=1e-2,
max_iters=50000)
ionosphere_results.append(round(acc, 2))
acc = cross_validation(5, X.to_numpy(), Processor.ToNumpyCol(Y), NaiveBayes())
ionosphere_results.append(round(acc, 2))
print(ionosphere_results)