In [2]:

    

import pandas as pd
%matplotlib inline


    

In [3]:

    

from sklearn import datasets
from pandas.tools.plotting import scatter_matrix


    

In [4]:

    

import matplotlib.pyplot as plt


    

In [5]:

    

iris = datasets.load_iris()


    

In [6]:

    

from sklearn.cross_validation import train_test_split


    

In [7]:

    

x = iris.data[:,2:]
y = iris.target


    

In [8]:

    

x_train, x_test, y_train, y_test = train_test_split(x,y,test_size=0.5,train_size=0.5)


    

In [9]:

    

from sklearn import tree


    

In [10]:

    

dt = tree.DecisionTreeClassifier()


    

In [11]:

    

dt = dt.fit(x_train,y_train)


    

In [12]:

    

from sklearn import metrics


    

In [13]:

    

def measure_performance(X,y,clf, show_accuracy=True, show_classification_report=True, show_confussion_matrix=True):
    y_pred=clf.predict(X)
    if show_accuracy:
        print("Accuracy:{0:.3f}".format(metrics.accuracy_score(y, y_pred)),"\n")
    if show_classification_report:
        print("Classification report")
        print(metrics.classification_report(y,y_pred),"\n")
    if show_confussion_matrix:
        print("Confusion matrix")
        print(metrics.confusion_matrix(y,y_pred),"\n")


    

In [14]:

    

measure_performance(x_train,y_train,dt)


    

    




Accuracy:1.000 

Classification report
             precision    recall  f1-score   support

          0       1.00      1.00      1.00        26
          1       1.00      1.00      1.00        21
          2       1.00      1.00      1.00        28

avg / total       1.00      1.00      1.00        75
 

Confusion matrix
[[26  0  0]
 [ 0 21  0]
 [ 0  0 28]] 

In [15]:

    

#They were 100% accurate for teh test data, which is good, but also has me suspicious.


    

In [16]:

    

x_train, x_test, y_train, y_test = train_test_split(x,y,test_size=0.25,train_size=0.75)


    

In [17]:

    

dt = tree.DecisionTreeClassifier()


    

In [18]:

    

dt = dt.fit(x_train,y_train)


    

In [19]:

    

def measure_performance(X,y,clf, show_accuracy=True, show_classification_report=True, show_confussion_matrix=True):
    y_pred=clf.predict(X)
    if show_accuracy:
        print("Accuracy:{0:.3f}".format(metrics.accuracy_score(y, y_pred)),"\n")
    if show_classification_report:
        print("Classification report")
        print(metrics.classification_report(y,y_pred),"\n")
    if show_confussion_matrix:
        print("Confusion matrix")
        print(metrics.confusion_matrix(y,y_pred),"\n")


    

In [20]:

    

measure_performance(x_train,y_train,dt)


    

    




Accuracy:0.991 

Classification report
             precision    recall  f1-score   support

          0       1.00      1.00      1.00        40
          1       1.00      0.97      0.99        40
          2       0.97      1.00      0.98        32

avg / total       0.99      0.99      0.99       112
 

Confusion matrix
[[40  0  0]
 [ 0 39  1]
 [ 0  0 32]] 

In [21]:

    

#They are slightly less accurate than before -- one of the results for the second group was placed in the third group.


    

In [22]:

    

brc = datasets.load_breast_cancer()


    

In [23]:

    

brc


    

    
Out[23]:





{'DESCR': 'Breast Cancer Wisconsin (Diagnostic) Database\n\nNotes\n-----\nData Set Characteristics:\n    :Number of Instances: 569\n\n    :Number of Attributes: 30 numeric, predictive attributes and the class\n\n    :Attribute Information:\n        - radius (mean of distances from center to points on the perimeter)\n        - texture (standard deviation of gray-scale values)\n        - perimeter\n        - area\n        - smoothness (local variation in radius lengths)\n        - compactness (perimeter^2 / area - 1.0)\n        - concavity (severity of concave portions of the contour)\n        - concave points (number of concave portions of the contour)\n        - symmetry \n        - fractal dimension ("coastline approximation" - 1)\n        \n        The mean, standard error, and "worst" or largest (mean of the three\n        largest values) of these features were computed for each image,\n        resulting in 30 features.  For instance, field 3 is Mean Radius, field\n        13 is Radius SE, field 23 is Worst Radius.\n        \n        - class:\n                - WDBC-Malignant\n                - WDBC-Benign\n\n    :Summary Statistics:\n\n    ===================================== ======= ========\n                                           Min     Max\n    ===================================== ======= ========\n    radius (mean):                         6.981   28.11\n    texture (mean):                        9.71    39.28\n    perimeter (mean):                      43.79   188.5\n    area (mean):                           143.5   2501.0\n    smoothness (mean):                     0.053   0.163\n    compactness (mean):                    0.019   0.345\n    concavity (mean):                      0.0     0.427\n    concave points (mean):                 0.0     0.201\n    symmetry (mean):                       0.106   0.304\n    fractal dimension (mean):              0.05    0.097\n    radius (standard error):               0.112   2.873\n    texture (standard error):              0.36    4.885\n    perimeter (standard error):            0.757   21.98\n    area (standard error):                 6.802   542.2\n    smoothness (standard error):           0.002   0.031\n    compactness (standard error):          0.002   0.135\n    concavity (standard error):            0.0     0.396\n    concave points (standard error):       0.0     0.053\n    symmetry (standard error):             0.008   0.079\n    fractal dimension (standard error):    0.001   0.03\n    radius (worst):                        7.93    36.04\n    texture (worst):                       12.02   49.54\n    perimeter (worst):                     50.41   251.2\n    area (worst):                          185.2   4254.0\n    smoothness (worst):                    0.071   0.223\n    compactness (worst):                   0.027   1.058\n    concavity (worst):                     0.0     1.252\n    concave points (worst):                0.0     0.291\n    symmetry (worst):                      0.156   0.664\n    fractal dimension (worst):             0.055   0.208\n    ===================================== ======= ========\n\n    :Missing Attribute Values: None\n\n    :Class Distribution: 212 - Malignant, 357 - Benign\n\n    :Creator:  Dr. William H. Wolberg, W. Nick Street, Olvi L. Mangasarian\n\n    :Donor: Nick Street\n\n    :Date: November, 1995\n\nThis is a copy of UCI ML Breast Cancer Wisconsin (Diagnostic) datasets.\nhttps://goo.gl/U2Uwz2\n\nFeatures are computed from a digitized image of a fine needle\naspirate (FNA) of a breast mass.  They describe\ncharacteristics of the cell nuclei present in the image.\nA few of the images can be found at\nhttp://www.cs.wisc.edu/~street/images/\n\nSeparating plane described above was obtained using\nMultisurface Method-Tree (MSM-T) [K. P. Bennett, "Decision Tree\nConstruction Via Linear Programming." Proceedings of the 4th\nMidwest Artificial Intelligence and Cognitive Science Society,\npp. 97-101, 1992], a classification method which uses linear\nprogramming to construct a decision tree.  Relevant features\nwere selected using an exhaustive search in the space of 1-4\nfeatures and 1-3 separating planes.\n\nThe actual linear program used to obtain the separating plane\nin the 3-dimensional space is that described in:\n[K. P. Bennett and O. L. Mangasarian: "Robust Linear\nProgramming Discrimination of Two Linearly Inseparable Sets",\nOptimization Methods and Software 1, 1992, 23-34].\n\nThis database is also available through the UW CS ftp server:\n\nftp ftp.cs.wisc.edu\ncd math-prog/cpo-dataset/machine-learn/WDBC/\n\nReferences\n----------\n   - W.N. Street, W.H. Wolberg and O.L. Mangasarian. Nuclear feature extraction \n     for breast tumor diagnosis. IS&T/SPIE 1993 International Symposium on \n     Electronic Imaging: Science and Technology, volume 1905, pages 861-870, \n     San Jose, CA, 1993. \n   - O.L. Mangasarian, W.N. Street and W.H. Wolberg. Breast cancer diagnosis and \n     prognosis via linear programming. Operations Research, 43(4), pages 570-577, \n     July-August 1995.\n   - W.H. Wolberg, W.N. Street, and O.L. Mangasarian. Machine learning techniques\n     to diagnose breast cancer from fine-needle aspirates. Cancer Letters 77 (1994) \n     163-171.\n',
 'data': array([[  1.79900000e+01,   1.03800000e+01,   1.22800000e+02, ...,
           2.65400000e-01,   4.60100000e-01,   1.18900000e-01],
        [  2.05700000e+01,   1.77700000e+01,   1.32900000e+02, ...,
           1.86000000e-01,   2.75000000e-01,   8.90200000e-02],
        [  1.96900000e+01,   2.12500000e+01,   1.30000000e+02, ...,
           2.43000000e-01,   3.61300000e-01,   8.75800000e-02],
        ..., 
        [  1.66000000e+01,   2.80800000e+01,   1.08300000e+02, ...,
           1.41800000e-01,   2.21800000e-01,   7.82000000e-02],
        [  2.06000000e+01,   2.93300000e+01,   1.40100000e+02, ...,
           2.65000000e-01,   4.08700000e-01,   1.24000000e-01],
        [  7.76000000e+00,   2.45400000e+01,   4.79200000e+01, ...,
           0.00000000e+00,   2.87100000e-01,   7.03900000e-02]]),
 'feature_names': array(['mean radius', 'mean texture', 'mean perimeter', 'mean area',
        'mean smoothness', 'mean compactness', 'mean concavity',
        'mean concave points', 'mean symmetry', 'mean fractal dimension',
        'radius error', 'texture error', 'perimeter error', 'area error',
        'smoothness error', 'compactness error', 'concavity error',
        'concave points error', 'symmetry error', 'fractal dimension error',
        'worst radius', 'worst texture', 'worst perimeter', 'worst area',
        'worst smoothness', 'worst compactness', 'worst concavity',
        'worst concave points', 'worst symmetry', 'worst fractal dimension'], 
       dtype='<U23'),
 'target': array([0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 0,
        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 0, 0, 0, 0, 0, 0, 0, 0,
        1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 0, 1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 1,
        1, 0, 1, 0, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 0, 0, 1, 1, 1, 0,
        1, 1, 0, 0, 1, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 1,
        1, 1, 0, 0, 0, 1, 0, 0, 1, 1, 1, 0, 0, 1, 0, 1, 0, 0, 1, 0, 0, 1, 1,
        0, 1, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1,
        0, 0, 1, 0, 1, 1, 0, 0, 1, 1, 0, 0, 1, 1, 1, 1, 0, 1, 1, 0, 0, 0, 1,
        0, 1, 0, 1, 1, 1, 0, 1, 1, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 0, 1, 0, 1,
        0, 1, 1, 0, 1, 0, 0, 0, 0, 1, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0,
        0, 1, 1, 0, 1, 1, 0, 0, 1, 0, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 0,
        0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 0, 1,
        1, 0, 1, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1,
        1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0,
        1, 0, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 1, 0, 1, 0, 1, 0, 1, 1, 1, 0, 1,
        1, 1, 1, 1, 1, 1, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 1,
        0, 0, 0, 1, 0, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 0, 1, 1, 1, 0, 1,
        1, 0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1,
        0, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 0, 0, 1, 0, 1,
        1, 1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 1, 0, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1,
        0, 0, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 1, 1, 1,
        1, 1, 1, 1, 0, 1, 0, 1, 1, 0, 1, 1, 1, 1, 1, 0, 0, 1, 0, 1, 0, 1, 1,
        1, 1, 1, 0, 1, 1, 0, 1, 0, 1, 0, 0, 1, 1, 1, 0, 1, 1, 1, 1, 1, 1, 1,
        1, 1, 1, 1, 0, 1, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1,
        1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 1]),
 'target_names': array(['malignant', 'benign'], 
       dtype='<U9')}

In [24]:

    

#We are trying to predict (whether the tumor is) malignant or benign
#The attributes we have are: radius, texture, perimeter, area, smoothness, compactness, concavity, concave points, symmetry, fractal dimension,


    

In [25]:

    

x_train, x_test, y_train, y_test = train_test_split(x,y,test_size=0.5,train_size=0.5)


    

In [26]:

    

dt = tree.DecisionTreeClassifier()


    

In [27]:

    

dt = dt.fit(x_train,y_train)


    

In [28]:

    

def measure_performance(X,y,clf, show_accuracy=True, show_classification_report=True, show_confussion_matrix=True):
    y_pred=clf.predict(X)
    if show_accuracy:
        print("Accuracy:{0:.3f}".format(metrics.accuracy_score(y, y_pred)),"\n")
    if show_classification_report:
        print("Classification report")
        print(metrics.classification_report(y,y_pred),"\n")
    if show_confussion_matrix:
        print("Confusion matrix")
        print(metrics.confusion_matrix(y,y_pred),"\n")


    

In [29]:

    

measure_performance(x_train,y_train,dt)


    

    




Accuracy:1.000 

Classification report
             precision    recall  f1-score   support

          0       1.00      1.00      1.00        21
          1       1.00      1.00      1.00        27
          2       1.00      1.00      1.00        27

avg / total       1.00      1.00      1.00        75
 

Confusion matrix
[[21  0  0]
 [ 0 27  0]
 [ 0  0 27]] 

In [32]:

    

x_train, x_test, y_train, y_test = train_test_split(x,y,test_size=0.25,train_size=0.75)


    

In [33]:

    

measure_performance(x_train,y_train,dt)


    

    




Accuracy:0.973 

Classification report
             precision    recall  f1-score   support

          0       1.00      1.00      1.00        38
          1       0.95      0.98      0.96        41
          2       0.97      0.94      0.95        33

avg / total       0.97      0.97      0.97       112
 

Confusion matrix
[[38  0  0]
 [ 0 40  1]
 [ 0  2 31]] 

In [ ]:

    

#Again, as with the irises, when larger training set was used accuracy decreased. 
#It actually decreased more as it went -- one was wrong from the second set, and two were wrong from the third set.