def test_linear_regression():
# Create a data frame
d = {'y' : pd.Series([95, 85, 80, 75, 70, 65, 60, 55, 50, 45], index=['1', '2', '3', '4', '5', '6', '7', '8', '9', '10']),
'x1' : pd.Series([85, 95, 70, 65, 70, 60, 64, 60, 51, 49], index=['1', '2', '3', '4', '5', '6', '7', '8', '9', '10']),
'x2' : pd.Series([10, 8.8, 8.4, 7.5, 7.4, 7.2, 7.0, 6.4, 5.3, 4], index=['1', '2', '3', '4', '5', '6', '7', '8', '9', '10']),
}
df = pd.DataFrame(d)
beta1, pvalues1 = linear_regression(df, 'y', 'x1')
beta2, pvalues2 = linear_regression(df, 'y', 'x2')
beta3, pvalues3 = linear_regression(df, 'y', 'x1', 'x2')
expected_beta1 = np.array([ 0.69128736, 1.00610931]) # Calculated by hands
expected_p1 = np.array([0.95669000991385234, 0.00082441892685309844]) # Calculated by hands
expected_beta2 = np.array([ 2.92830189, 9.03773585]) # Calculated by hands
expected_p2 = [0.64660353670191761, 1.2010523101013017e-05] # Calculated by hands
expected_beta3 = np.array([-0.01554384, 0.20355359, 7.55525124]) # Calculated by hands
expected_p3 = np.array([0.99826544217405555, 0.37237722050579208, 0.0049816157477362418]) # Calculated by hands
assert_almost_equal(expected_beta1, beta1)
assert_almost_equal(expected_p1, pvalues1)
assert_almost_equal(expected_beta2, beta2)
assert_almost_equal(expected_p2, pvalues2)
assert_almost_equal(expected_beta3, beta3)
assert_almost_equal(expected_p3, pvalues3)
def test_linear_regression():
# Create a data frame
d = {
"y": pd.Series(
[95, 85, 80, 75, 70, 65, 60, 55, 50, 45], index=["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"]
),
"x1": pd.Series(
[85, 95, 70, 65, 70, 60, 64, 60, 51, 49], index=["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"]
),
"x2": pd.Series(
[10, 8.8, 8.4, 7.5, 7.4, 7.2, 7.0, 6.4, 5.3, 4], index=["1", "2", "3", "4", "5", "6", "7", "8", "9", "10"]
),
}
df = pd.DataFrame(d)
beta1, pvalues1 = linear_regression(df, "y", "x1")
beta2, pvalues2 = linear_regression(df, "y", "x2")
beta3, pvalues3 = linear_regression(df, "y", "x1", "x2")
expected_beta1 = np.array([0.69128736, 1.00610931]) # Calculated by hands
expected_p1 = np.array([0.95669000991385234, 0.00082441892685309844]) # Calculated by hands
expected_beta2 = np.array([2.92830189, 9.03773585]) # Calculated by hands
expected_p2 = [0.64660353670191761, 1.2010523101013017e-05] # Calculated by hands
expected_beta3 = np.array([-0.01554384, 0.20355359, 7.55525124]) # Calculated by hands
expected_p3 = np.array([0.99826544217405555, 0.37237722050579208, 0.0049816157477362418]) # Calculated by hands
assert_almost_equal(expected_beta1, beta1)
assert_almost_equal(expected_p1, pvalues1)
assert_almost_equal(expected_beta2, beta2)
assert_almost_equal(expected_p2, pvalues2)
assert_almost_equal(expected_beta3, beta3)
assert_almost_equal(expected_p3, pvalues3)
def run_linear(train, test):
print("RUNNING LINEAR")
#convert it to another dataframe with ages
train = add_information(train)
train.to_csv("test.csv")
model = linear_regression(train)
test = add_information(test)
return evaluate_linear_regression(model, test)
def brute_force(data):
for col1 in data.columns:
X = data[col1]
if type(X[1]) != type(str(
)): # this is dirty but i don't have internet (pls don't judge ^^ )
for col2 in data.columns:
y = data[col2]
if type(y[1]) != type('str') and (col1 != col2):
y_p = linear_regression(X, y)
plot_graph(data, X, y, y_p)
def performing_algorithm(X, y, X_test):
"""
:param X: Matrix
:param y: Matrix
:param X_test: Matrix
:return: Prediction of chosen algorithm
"""
if args.algorithm == "linear_regression":
return linear_regression(X, y, X_test)
elif args.algorithm == "decision_tree":
return decision_tree(X, y, X_test)
elif args.algorithm == "SVM":
return SVM(X, y, X_test)
def choose_function(data):
choice = int(
input("choose a function #0 Linear regression, #1 bruteforce :"))
if choice == 0:
print("linear regression choosed.")
X, y = columns_selection(data)
y_p = linear_regression(X, y)
plot_graph(data, X, y, y_p)
elif choice == 1:
print("bruteforce choosed.")
brute_force(data)
else:
choose_function(data)
def plot_correlation(
folder,
lambda_alpha):
file_path = os.path.join(folder,'gompertz','individualParameters','estimatedIndividualParameters.txt')
df = pd.read_csv(file_path, sep = ',')
alpha = df['alpha0_mode']
beta = df['beta_mode']
lr = linear_regression(beta.values, alpha.values, 1)
fig, ax, fig_text = plot_regression_line(beta, alpha, lr.params, np.min([lr.rsquared, 0.99]), np.max([1e-5,np.min(lr.pvalues)]))
ax.plot([beta.min(),beta.max()], lambda_alpha*np.ones(2),'k:', label = '$\lambda$ = '+np.str(lambda_alpha))
ax.legend(loc=4, fontsize=13)
figname=os.path.join(folder, 'correlation.pdf')
fig.savefig(figname, dpi = 1000, format = 'pdf', bbox_inches='tight')
fig_text.savefig(os.path.join(folder, 'correlation_box.pdf'), dpi = 1000, format = 'pdf', bbox_inches='tight')
def age_prediction(matrix, column):
"""
:param matrix: Matrix
:param column: Int
:return: Filled age column with predicted age
"""
train_age = []
test_age = []
index = []
for line in range(len(matrix)):
if pd.isnull(matrix[line][column]):
test_age.append(matrix[line])
index.append(line)
else:
train_age.append(matrix[line])
test_age = np.array(test_age)
train_age = np.array(train_age)
X_age = np.delete(train_age, [column], 1)
X_age = X_age.astype(np.float)
y_age = train_age[:, column]
y_age = y_age.astype(np.int)
X_test_age = np.delete(test_age, [column], 1)
X_test_age = X_test_age.astype(np.float)
predicted_age = linear_regression(X_age, y_age, X_test_age)
for line in range(len(predicted_age)):
if predicted_age[line] < 0:
predicted_age[line] = 1
var = 0
for line in range(len(matrix)):
if pd.isnull(matrix[line][column]):
matrix[line][column] = predicted_age[var]
var += 1
data_each.append(load_data(i, data_dir)) #data_each.append(load_data(i)) for i in range(len(all_subjects)): data_each[i]['ratio'] = data_each[i]['gain'] / data_each[i]['loss'] ############################## # Peform linear regression # ############################## data = all_data # Run the linear_regression function to get the summary beta1, pvalues1 = linear_regression(data, 'RT', 'gain', 'loss') beta2, pvalues2 = linear_regression(data, 'RT', 'ratio') beta3, pvalues3 = linear_regression(data, 'RT', 'diff') ####################### # Plot # ####################### # PLot the simple regression # Since the ratio is the most significant predictor y = data['RT']
############### Name: Shubham Pareek ############ ############### UBID: spareek ############ from logistic_regression import * from linear_regression import * from neural_network import * from preprocessing import * X1, y1 = get_feature_matrix(data='hod', method='concatenate') X2, y2 = get_feature_matrix(data='hod', method='subtract') X3, y3 = get_feature_matrix(data='gsc', method='concatenate') X4, y4 = get_feature_matrix(data='gsc', method='subtract') logistic_regression(X1, y1) logistic_regression(X2, y2) logistic_regression(X3, y3) logistic_regression(X4, y4) linear_regression(X1, y1) linear_regression(X2, y2) linear_regression(X3, y3) linear_regression(X4, y4) neural_network(X1, y1) neural_network(X2, y2) neural_network(X3, y3) neural_network(X4, y4)
data_each = [] for i in all_subjects: data_each.append(load_data(i, data_dir)) for i in range(len(all_subjects)): data_each[i]['ratio'] = data_each[i]['gain'] / data_each[i]['loss'] ####################### # Peform regression # ####################### data = all_data # Run the linear_regression function to get the summary linear_regression(data, 'RT', 'gain', 'loss') linear_regression(data, 'RT', 'ratio') linear_regression(data, 'RT', 'diff') ####################### # Plot # #######################
np.random.seed(42)
x = dataset.data
y = dataset.target
indices = np.random.permutation(len(x))
test_size = 100
x_train = x[indices[:-test_size]]
y_train = y[indices[:-test_size]]
x_test = x[indices[-test_size:]]
y_test = y[indices[-test_size:]]
regr = linear_regression()
regr.fit(x_train, y_train)
print("Coeffs: ", regr.beta[1:])
print("Intercept: ", regr.beta[0])
print("R2: ", regr.score(x_test, y_test))
train_pred = regr.predict(x_train)
test_pred = regr.predict(x_test)
min_val = min(min(train_pred), min(test_pred))
max_val = max(max(train_pred), max(test_pred))
# y_pred = 10, y = 12
# -2
plt.scatter(train_pred, train_pred - y_train, color="blue", s=40)
