def plot_epsilon_graph():
"""
plot an epsilon graph
@return: list of cluster labels
"""
warnings.filterwarnings("ignore")
df = pd.read_csv('houshold2007.csv')
df = df[df['Date'].apply(lambda x: x.endswith('1/2007'))]
preprocessing(df)
attributes = df.columns.tolist()
data = df.to_numpy()
nbrs = NearestNeighbors(n_neighbors=get_min_pts(attributes),
algorithm='auto').fit(data)
distances, indices = nbrs.kneighbors(data)
temp = []
for i in range(len(distances)):
temp.append(distances[i][len(distances[i]) - 1])
temp.sort(reverse=True)
plt.plot(temp)
plt.grid(True)
plt.xaxis('test')
plt.title('Epsilon Threshold')
plt.savefig('eps_plot.png')
plt.show()
def overUnderScatter(
): # this returns a scatter plot of an over/under simulation
n = 500
x = np.random.rand(n)
y = winList
colors = (0, 0, 0)
plt.scatter(x, y, c=colors, alpha=0.5)
plt.set(title="Over Under Sim Returns")
plt.ylabel('Profit')
plt.xaxis('U.S. Dollar (Thousands)')
plt.show()
def plot_scatter_of_body(dataframe, feature_1="Body", feature_2="Score"):
fig, ax = plt.subplots(1, figsize=(30, 6))
dataframe["Title_len"] = dataframe[feature_1].str.split().str.len()
dataframe = dataframe.groupby("Title_len")[feature_2].mean().reset_index()
x = dataframe["Title_len"]
y = dataframe["Score"]
sns.scatterplot(x=x, y=y, data=dataframe, legend="brief", ax=ax)
plt.title("Average Upvote by Question Body Length")
plt.yaxis("Average Upvote")
plt.xaxis("Question Body Length")
plt.plot()
rangeN = range(1, numberSubIntervalls)
t, w = np.polynomial.legendre.leggauss(
order) # t in [-1,1], w weight, ordre 50
plt.figure()
for epsilon in [1, 2, 3]:
print epsilon
a1 = np.array([-epsilon, -1.0])
b1 = np.array([-epsilon, +1.0])
a2 = np.array([epsilon, -1.0])
b2 = np.array([epsilon, +1.0])
int1 = [a1, b1]
int2 = [a2, b2]
errSing = np.array([])
for n in rangeN:
errSing = np.append(errSing, calculateError(int1, int2, n, order, t,
w))
#print "coucou", calculateInt(int1, int2, order, t, w)
print(np.log(errSing))
plt.plot(np.log(rangeN), np.log(errSing), label=epsilon)
plt.hold(True)
plt.legend()
plt.title("singular function for different eps")
plt.yaxis("error")
plt.xaxis("ln(number sub intervalls)")
feed_dict={
x: trainX[start:end],
y: trainY[start:end]
})
if ((i + 1) % 7 == 0):
[mseloss, crossError] = sess.run([loss, crossLoss],
feed_dict={
x: trainX[start:end],
y: trainY[start:end]
})
mseValues.append(mseloss)
crossValues.append(crossError)
[mseClass, crossClass] = sess.run([mseClassError, crossClassError],
feed_dict={
x: trainX[start:end],
y: trainY[start:end]
})
mseClassValues.append(mseClass)
crossClassValues.append(1 - crossClass)
plt.title("Accuracy of Linear vs Logistic")
plt.xaxis("epochs")
plt.yaxis("accuracy")
plt.plot(mseClassValues, label="linear")
plt.plot(crossClassValues, label="logistic")
plt.legend()
plt.show()
print(
i,
sqrt(np.abs(np.mean(test_scaled[:, 1, i]**2 - predictions[:, i]**2))) /
sqrt(np.mean(test_scaled[:, 1, i]**2)))
out = 13
#pyplot.plot(series[out])
#pyplot.show()
#pyplot.plot(train_scaled_x,train_scaled[:,0,out])
#pyplot.plot(test_scaled_x,test_scaled[:,0,out])
#pyplot.plot(test_scaled_x,predictions[:,out])
#pyplot.show()
pyplot.plot(test_scaled_x, test_scaled[:, 0, out])
pyplot.plot(test_scaled_x, predictions[:, out])
pyplot.xaxis('time (minutes)')
pyplot.yaxis('B (nT)')
pyplot.show()
#pyplot.plot(nrmse)
#pyplot.show()
# track loss vs records
# array output/prediction
# null hypothesis
# check noise in data. Take diff with stencil?
# improve data saving technique
gray.shape[::-1], None,
None)
return mtx, dist
mtx, dist = camera_calibration()
#img = mpimg.imread('camera_cal/calibration1.jpg')
img = mpimg.imread('test_images/test1.jpg')
undist = cv2.undistort(img, mtx, dist, None, mtx)
fig = plt.figure(figsize=(20, 10))
fig.add_subplot(1, 2, 1)
plt.imshow(img)
plt.xaxis()
fig.add_subplot(1, 2, 2)
plt.imshow(undist)
#cv2.destroyAllWindows()
#fig.savefig('output_images/chessboard_undistort.jpg')
#fig.savefig('output_images/test1_undistort.jpg')
# ## Next I apply a color transform consisting of a combination of a threshhold to the gradient in the x-direction and to the s-channel
# In[3]:
# Color transform
img = mpimg.imread('test_images/test1.jpg')
data = pd.read_csv('Position_Salaries.csv')
x = data.iloc[:, 1:2].values
y = data.iloc[:, 2].values
from sklearn.preprocessing import PolynomialFeatures
p = PolynomialFeatures(degree=4)
x_poly = p.fit_transform(x)
from sklearn.linear_model import LinearRegression
r = LinearRegression()
r.fit(x_poly, y)
plt.scatter(x, y, color='red')
plt.plot(x, r.predict(p.fit_transform(x)), color='blue')
plt.title("Position v/s salaries")
plt.xaxis("positions")
plt.yaxis("salaries")
plt.show()
#making the curve more smooother
x_grid = np.arange(min(x), max(x), 0.1)
x_grid = x_grid.reshape(len(x_grid), 1)
plt.scatter(x, y, color='red')
plt.plot(x_grid, r.predict(p.fit_transform(x_grid)), color='blue')
plt.title("Position v/s salaries")
plt.xlabel("positions")
plt.ylabel("salaries")
plt.show()
x_ans = np.array(6.5)