Updated

README.md

@@ -0,0 +1,16 @@
+# machinelearningregression
+boilerplate code, scripts, modules, data for Regression Analysis workshop
+
+Before starting check that your working environment is ready running the `runme.py`:
+
+```
+$ python runme.py
+```
+
+If the following message appears:
+
+```
+Good to go, all packages installed ok, ready to code.
+```
+
+You're ready fot the workshop!

functions.py

@@ -1,4 +1,4 @@
-#cca regression functions
+# Advanced functions
 from sklearn.pipeline import make_pipeline
 from sklearn.linear_model import LinearRegression
 from sklearn.preprocessing import StandardScaler, PolynomialFeatures
@@ -23,23 +23,20 @@ def model_plot_3d(ax, model, x1, x2):
     positions = np.vstack([X1.ravel(), X2.ravel()]).T
     y = model.predict(positions)
     Y = y.reshape(X1.shape)
-    ax.plot_surface(X1, X2, Y, alpha = 0.2 , cmap = 'jet')
-
-def save_figure(name, i):
-    plt.savefig('../figures/'+name+str(i), bbox_inches='tight')
-    plt.gcf().clear()
+    ax.plot_surface(X1, X2, Y, alpha = 0.2 , cmap = 'jet',\
+                    linewidth=0.5, rstride=1, cstride=1, shade=True)
 
 def PolynomialRegression(degree = 1):
-    return make_pipeline(PolynomialFeatures(degree = degree), LinearRegression())
+    return make_pipeline(PolynomialFeatures(degree = degree,\
+                            include_bias = False), LinearRegression())
 
 def PolynomialRidge(degree = 1, alpha = 1):
-    return make_pipeline(PolynomialFeatures(degree = degree), StandardScaler(), Ridge(alpha = alpha))
+    return make_pipeline(PolynomialFeatures(degree = degree,\
+                            include_bias = False), StandardScaler(), Ridge(alpha = alpha))
 
 def PolynomialLasso(degree = 1, alpha = 1):
-    return make_pipeline(PolynomialFeatures(degree = degree), StandardScaler(), Lasso(alpha = alpha))
-
-
-
+    return make_pipeline(PolynomialFeatures(degree = degree,\
+                            include_bias = False), StandardScaler(), Lasso(alpha = alpha))
 
 def polynomial_residual(degree, X, y):
     polynomial_regression= PolynomialRegression(degree = degree)
@@ -48,54 +45,11 @@ def polynomial_residual(degree, X, y):
     mae = mean_absolute_error(y, y_pred)
     return mae
 
-
 def organize_data(to_forecast, window, horizon):
-    """
-     Input:
-      to_forecast, univariate time series organized as numpy array
-      window, number of items to use in the forecast window
-      horizon, horizon of the forecast
-     Output:
-      X, a matrix where each row contains a forecast window
-      y, the target values for each row of X
-    """
     shape = to_forecast.shape[:-1] + (to_forecast.shape[-1] - window + 1, window)
     strides = to_forecast.strides + (to_forecast.strides[-1],)
     X = np.lib.stride_tricks.as_strided(to_forecast,
                                         shape=shape,
                                         strides=strides)
     y = np.array([X[i+horizon][-1] for i in range(len(X)-horizon)])
     return X[:-horizon], y
-
-
-class NonLinearRegression(object):
-
-    def __init__(self, fun):
-        self.fun = fun
-        self._find_n_params()
-
-    def fit(self, X, y):
-        self.params = basinhopping(lambda p: self._get_residual(X,y,p), np.random.randn(self._n_params), niter = 10000, niter_success = 50)['x']
-
-    def predict(self, X):
-        return self.fun(X.flatten(), self.params)
-
-    def get_params(self, deep=True):
-        return {"fun": self.fun}
-
-    def set_params(self, **parameters):
-        for parameter, value in parameters.items():
-            self.setattr(parameter, value)
-        return self
-
-    def _get_residual(self, X, y, p):
-        return np.mean((self.fun(X.flatten(), p)-y)**2)
-
-    def _find_n_params(self):
-        for n_params in range(1,100)[::-1]:
-            try:
-                self.fun(1,np.random.rand(n_params))
-            except:
-                n_params += 1
-                break
-        self._n_params = n_params

module3.py

@@ -1,16 +1,10 @@
-from functions import save_figure
-from os.path import splitext
-figname = splitext(__file__)[0]+'_'
-ifig = 0
+# Module 3: Linear regression
 
+# New imports
 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 from sklearn.linear_model import LinearRegression
+from sklearn import metrics
 
-bikes_df = pd.read_csv('./data/bikes_subsampled.csv')
-
-# We select the variables temperature and bikes_count
-
-temperature = bikes_df['temperature'].values
-bikes_count = bikes_df['count'].values
+# Code after this

module4.py

@@ -1,27 +1,18 @@
-from functions import save_figure
-from os.path import splitext
-figname = splitext(__file__)[0]+'_'
-ifig = 0
-
-################################################################################
-################################### MODULE 4 ###################################
-##################### Multiple variables linear regression #####################
-################################################################################
+# Module 4: Multiple and polynomial regression
 
+# Previous imports
 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 from sklearn.linear_model import LinearRegression
-from mpl_toolkits.mplot3d.axes3d import Axes3D
-from sklearn.preprocessing import PolynomialFeatures
-
-from functions import PolynomialRegression, model_plot_3d
 
+# New imports
+from functions import PolynomialRegression
+from mpl_toolkits.mplot3d.axes3d import Axes3D
+from functions import model_plot_3d
+from functions import polynomial_residual
 
+# Load dataset
 bikes_df = pd.read_csv('./data/bikes_subsampled.csv')
 
-# Learning activity 1: Fit a model of 2 variables and plot the model
-
-features = ['temperature','humidity']
-X = bikes_df[features].values
-y = bikes_df['count'].values
+# Code after this

module5.py

@@ -1,23 +1,19 @@
-from functions import save_figure
-from os.path import splitext
-figname = splitext(__file__)[0]+'_'
-ifig = 0
-
-################################################################################
-################################### MODULE 5 ###################################
-############################# Model evaluation #################################
-################################################################################
-# To be separated in a unique file #
+# Module 5: Evaluating model performance
 
+# Previous imports
 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
 from sklearn.linear_model import LinearRegression
 from functions import PolynomialRegression
-from sklearn.preprocessing import PolynomialFeatures
+
+# New imports
 from sklearn.cross_validation import train_test_split, cross_val_score
 from sklearn.metrics import mean_absolute_error
 
+# Load dataset
 bikes_df = pd.read_csv('./data/bikes_subsampled.csv')
 temperature = bikes_df[['temperature']].values
 bikes_count = bikes_df['count'].values
+
+# Code after this

module6.py

@@ -1,26 +1,18 @@
-from functions import save_figure
-from os.path import splitext
-figname = splitext(__file__)[0]+'_'
-ifig = 0
+# Module 6: Avoid overfitting with regularisation
 
-
-################################################################################
-################################### MODULE 6 ###################################
-############################# Regularisation ###################################
-################################################################################
-
-
-# Learning activity 2: Ridge and Lasso regularisations
+# Previous imports
 import pandas as pd
 import numpy as np
 import matplotlib.pyplot as plt
-from sklearn.pipeline import make_pipeline
-from sklearn.preprocessing import StandardScaler, PolynomialFeatures
-from sklearn.linear_model import Ridge, Lasso
+from sklearn.linear_model import LinearRegression
+from functions import PolynomialRegression
+from sklearn.cross_validation import cross_val_score
 
-from functions import PolynomialRidge, PolynomialLasso, PolynomialRegression
+# New imports
+from functions import PolynomialRidge, PolynomialLasso
+from sklearn.grid_search import GridSearchCV
 
+# Load dataset
 bikes_df = pd.read_csv('./data/bikes.csv')
-features = ['temperature','humidity','windspeed']
-X = bikes_df[features].values
-y = bikes_df['count'].values
+
+# Code after this

module7.py

@@ -1,30 +1,17 @@
-from functions import save_figure
-from os.path import splitext
-figname = splitext(__file__)[0]+'_'
-ifig = 0
+# Module 7: Predict the future with autoregression
 
-################################################################################
-################################### MODULE 7 ###################################
-############################# Advanced fitting methods #########################
-################################################################################
-
-#Learning activity 1: use any sklearn model
-
-from sklearn.tree import DecisionTreeRegressor
-from sklearn.cross_validation import cross_val_score
-from sklearn.linear_model import LinearRegression
-from sklearn import metrics
-from sklearn.ensemble import RandomForestRegressor
-from sklearn.neighbors import KNeighborsRegressor
-from sklearn.svm import SVR
-import numpy as np
+# Previous imports
 import pandas as pd
+import numpy as np
 import matplotlib.pyplot as plt
+from sklearn.linear_model import LinearRegression
+from sklearn.metrics import mean_absolute_error
 
-from functions import NonLinearRegression, organize_data
-
+# New imports
+from functions import organize_data
 
-#Learning activity 2: Custom nonlinear regression
+# Load dataset
 bikes_df = pd.read_csv('./data/bikes.csv')
-temperature = bikes_df[['temperature']].values
 bikes = bikes_df['count'].values
+
+# Code after this

runme.py

@@ -1,11 +1,7 @@
-import pandas
-import matplotlib.pyplot as plt
-from sklearn import preprocessing
-from sklearn.cross_validation import train_test_split
-from sklearn import metrics
-import functions
-from sklearn import linear_model
-from sklearn import cross_validation
-from sklearn.grid_search import GridSearchCV
+import module3
+import module4
+import module5
+import module6
+import module7
 
 print "Good to go, all packages installed ok, ready to code."