cwd = os.getcwd()
cwd = cwd[0:len(cwd) - 11]
os.chdir(cwd)
from import_data import get, coeff
from sklearn.model_selection import train_test_split, LeaveOneOut, KFold
from sklearn.neighbors import KNeighborsClassifier
from sklearn.metrics import confusion_matrix
from sklearn import metrics
import matplotlib.pyplot as plt
import numpy as np
#%% mostrar el rendimiento del sistema
if not os.path.exists('resultados/knn'):
os.makedirs('resultados/knn')
data, target = get('data', random=2)
for i in range(3):
random = i
X_train, X_test, y_train, y_test = get("cross", random)
k_range = range(1, 13, 2)
acc = []
acc1 = []
print(i)
sens = []
pre = []
for j in k_range:
knn = KNeighborsClassifier(n_neighbors=j)
knn.fit(X_train, y_train)
# dn = pd.concat([ds,dx], axis = 1)
# dn = pd.concat([dn,dw], axis = 1)
# [f, c] = dn.shape
# c = c - 1
# data = np.zeros((f,c))
# target = dn.get('label').values
# columns = dn.columns
# for i in range(len(columns) - 1):
# data[:,i] = dn.get(columns[i]).values
knn = KNeighborsClassifier(n_neighbors=5)
X_train, X_test, y_train, y_test = get('cross', 2, pca=True)
knn.fit(X_train, y_train)
print(knn.score(X_test, y_test))
X_train, X_test, y_train, y_test = get('cross', 2, pca=False)
knn.fit(X_train, y_train)
print(knn.score(X_test, y_test))
"""
test 3
X_train, X_test, y_train, y_test = get('cross', random = 2)
def app(**kwargs):
start_args.update(kwargs)
view_mode = start_args['view_mode']
data_mode = start_args['data_mode']
data_length = start_args['data_length']
has_bool = start_args['has_bool']
if data_mode == enums.DataMode.TEST and data_length is None:
print("Please specify data_length!")
return
if enums.ViewMode.EEL in view_mode:
import eel # pylint: disable=import-outside-toplevel
import eel_func # pylint: disable=import-outside-toplevel
eel.init("view")
eel.start("index.html", port=0, block=False, close_callback=eel_func.close)
eel.sleep(start_args['eel_start_delay'])
if enums.ViewMode.GRAPH in view_mode:
import plot_func # pylint: disable=import-outside-toplevel
plot_func.init()
if enums.ViewMode.CREATE_DATA in view_mode:
import create_data # pylint: disable=import-outside-toplevel
f = create_data.init(start_args['create_file_name'])
if data_mode == enums.DataMode.SERIAL:
import serial_func # pylint: disable=import-outside-toplevel
ser = serial_func.start(start_args['port'], view_mode)
if ser is None:
sys.exit()
elif data_mode == enums.DataMode.TEST:
import import_data # pylint: disable=import-outside-toplevel
init_data = import_data.init(start_args['import_file_name'])
j = 0
if data_mode == enums.DataMode.SERIAL:
data = serial_func.get_data(ser)
if data_length is None:
data_length = (len(data) - 1) / 2 if has_bool else len(data) - 1
while len(data) == 0:
print(data)
data = serial_func.get_data(ser)
elif data_mode == enums.DataMode.TEST:
data = import_data.get(data_length, j, init_data, has_bool)
j += 1
t, ys, tInt = np_func.init(data, start_args['size'], data_length)
old_result = None
if has_bool:
bools = np_func.bool_init(data_length, start_args['size'])
bools = np_func.set_bool_data(bools, data, data_length)
data = data[:(data_length + 1)]
while True:
try:
if data_mode == enums.DataMode.SERIAL:
data = serial_func.get_data(ser)
elif data_mode == enums.DataMode.TEST:
data = import_data.get(data_length, j, init_data, has_bool)
j += 1
if has_bool:
bools = np_func.set_bool_data(bools, data, data_length)
data = data[:(data_length + 1)]
if enums.ViewMode.TERMINAL in view_mode:
print(data)
if enums.ViewMode.CREATE_DATA in view_mode:
create_data.write(f, data)
if len(data) != data_length + 1:
continue
t, ys = np_func.set_data(data, t, ys, tInt, data_length)
if enums.ViewMode.GRAPH in view_mode:
plot_func.update(t, ys, data_length)
if not enums.ViewMode.CREATE_DATA in view_mode:
if has_bool:
result, draw = start_args['check_function'](t, ys, data_length, bools)
else:
result, draw = start_args['check_function'](t, ys, data_length)
if enums.ViewMode.TERMINAL in view_mode:
print("check result is ")
print(result)
if not enums.ViewMode.CREATE_DATA in view_mode:
if enums.ViewMode.EEL in view_mode:
if old_result is not None:
if not np_func.equal(result, old_result):
eel_func.render(result, data_length)
elif draw:
eel_func.render(result, data_length)
old_result = np_func.copy(result)
except KeyboardInterrupt:
# FIX ME: 一発で抜けてくれない。matplotlibのtkinterが悪そう。
if data_mode == enums.DataMode.SERIAL:
ser.close()
if enums.ViewMode.EEL in view_mode:
eel.close_page() # pylint: disable=no-member
if enums.ViewMode.CREATE_DATA in view_mode:
f.close()
break
# -*- coding: utf-8 -*-
"""
Created on Mon Jun 15 17:20:44 2020
@author: Matador
"""
import os
cwd = os.getcwd()
cwd = cwd[0:len(cwd) - 11]
os.chdir(cwd)
from import_data import get
from sklearn.svm import SVC
from sklearn.model_selection import train_test_split
from sklearn.metrics import confusion_matrix
import matplotlib.pyplot as plt
data, target = get('data', 2, pca=True)
X_train, X_test, y_train, y_test = train_test_split(data,
target,
random_state=0,
test_size=0.2)
svm = SVC(degree=5).fit(X_train, y_train)
y_pred = svm.predict(X_test)
score = []
matrix_c = confusion_matrix(y_test, y_pred)
score.append(svm.score(X_test, y_test))
axes = ax[i]
axes.set_title(titles[i])
axes.plot(time,variables[names[i]])
axes.grid()
axes.set_ylabel('voltage (mv)')
axes.set_xlabel('tiempo (seg)')
if not os.path.exists("info/orders"):
os.makedirs("info/orders", exist_ok = True)
fig.savefig('info/orders/tratamiento_de_la_señal.png')
#%% generar la matriz de correlacion
data = get('Dataframe' , random = 2, pca = False)
plot_corr(data)
plt.savefig('info/orders/matriz_correlacion.png')
plt.close('all')
#%% generar grafica de dispersion
X_train, X_test, y_train, y_test = get('cross', random = 2, pca = False)
iris_dataframe = pd.DataFrame(X_train)
grr = pd.plotting.scatter_matrix(iris_dataframe, c=y_train, figsize=(15, 15), marker='o',
hist_kwds={'bins': 20}, s=60, alpha=.8)
plt.savefig('info/orders/diagrama de dispersion.png')