Week 6 Peer-graded Assignment: The best classifier | Robotics With Python
Now that you have been equipped with the skills to use different Machine Learning algorithms, over the course of five weeks, you will have the opportunity to practice and apply it on a dataset. In this project, you will complete a notebook where you will build a classifier to predict whether a loan case will be paid off or not.
You load deoxyadenosine monophosphate historical dataset from previous loan application, clean the data, and apply different classification algorithm on the datum. You be have a bun in the oven to use the succeed algorithm to build your model :
k-Nearest neighbor
decisiveness tree
subscribe vector machine
logistic regression
The resultant role be reported vitamin a the accuracy of each classifier, use the take after prosody when these constitute applicable :
You load deoxyadenosine monophosphate historical dataset from previous loan application, clean the data, and apply different classification algorithm on the datum. You be have a bun in the oven to use the succeed algorithm to build your model :
k-Nearest neighbor
decisiveness tree
subscribe vector machine
logistic regression
The resultant role be reported vitamin a the accuracy of each classifier, use the take after prosody when these constitute applicable :
Jaccard index
F1-score
LogLoass
review criterialess
This final project will equal grade aside your peer world health organization be complete this course during the like school term. This plan be worth twenty-five stigmatize of your entire grade, broken down a stick to :build exemplary practice KNN, find the good kilobyte and accuracy evaluation ( seven mark )
build model use decision tree, find the good kilobyte and accuracy evaluation ( six tag )
build model use SVM, recover the well k and accuracy evaluation ( six punctuate )
build up exemplary use logistic regression, rule the well thousand and accuracy evaluation ( six stigmatize )Read more : IBM System/360 – Wikipedia
Read more : Máy tính cá nhân IBM – Wikipedia tiếng Việt
visualize url GitHub depository – hypertext transfer protocol : //github.com/AbhishekTyagi404/IBM-ML-Loan-Prediction/tree/master/Coursera % 20ML % 20Graded % 20Assignment
Code-
In this notebook we try to practice all the classification algorithms that we learned in this course.
We load a dataset using Pandas library, and apply the following algorithms, and find the best one for this specific dataset by accuracy evaluation methods.
Lets first load required libraries:
In [5]:
import itertools
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.ticker import NullFormatter
import pandas as pd
import numpy as np
import matplotlib.ticker as ticker
from sklearn import preprocessing
%matplotlib inline
About dataset
This dataset is about past loans. The Loan_train.csv data set includes details of 346 customers whose loan are already paid off or defaulted. It includes following fields:
Field Description
Loan_status Whether a loan is paid off on in collection
Principal Basic principal loan amount at the
Terms Origination terms which can be weekly (7 days), biweekly, and monthly payoff schedule
Effective_date When the loan got originated and took effects
Due_date Since it’s one-time payoff schedule, each loan has one single due date
Age Age of applicant
Education Education of applicant
Gender The gender of applicant
Lets download the dataset
In [6]:
from __future__ import print_function
import os
data_path = ['loan_train']
print (data_path)
['loan_train']
Load Data From CSV File
In [7]:
df = pd.read_csv('loan_train.csv')
df.head()
Out[7]:
Unnamed: 0 Unnamed: 0.1 loan_status Principal terms effective_date due_date age education Gender
0 0 0 PAIDOFF 1000 30 9/8/2016 10/7/2016 45 High School or Below male
1 2 2 PAIDOFF 1000 30 9/8/2016 10/7/2016 33 Bechalor female
2 3 3 PAIDOFF 1000 15 9/8/2016 9/22/2016 27 college male
3 4 4 PAIDOFF 1000 30 9/9/2016 10/8/2016 28 college female
4 6 6 PAIDOFF 1000 30 9/9/2016 10/8/2016 29 college male
In [8]:
df.shape
Out[8]:
(346, 10)
Convert to date time object
In [9]:
df['due_date'] = pd.to_datetime(df['due_date'])
df['effective_date'] = pd.to_datetime(df['effective_date'])
df.head()
Out[9]:
Unnamed: 0 Unnamed: 0.1 loan_status Principal terms effective_date due_date age education Gender
0 0 0 PAIDOFF 1000 30 2016-09-08 2016-10-07 45 High School or Below male
1 2 2 PAIDOFF 1000 30 2016-09-08 2016-10-07 33 Bechalor female
2 3 3 PAIDOFF 1000 15 2016-09-08 2016-09-22 27 college male
3 4 4 PAIDOFF 1000 30 2016-09-09 2016-10-08 28 college female
4 6 6 PAIDOFF 1000 30 2016-09-09 2016-10-08 29 college male
Data visualization and pre-processing
Let’s see how many of each class is in our data set
In [10]:
df['loan_status'].value_counts()
Out[10]:
PAIDOFF 260
COLLECTION 86
Name: loan_status, dtype: int64
260 people have paid off the loan on time while 86 have gone into collection
Lets plot some columns to underestand data better:
In [11]:
# notice: installing seaborn might takes a few minutes
#!conda install -c anaconda seaborn -y
In [12]:
import seaborn as sns
bins = np.linspace(df.Principal.min(), df.Principal.max(), 10)
g = sns.FacetGrid(df, col="Gender", hue="loan_status", palette="Set1", col_wrap=2)
g.map(plt.hist, 'Principal', bins=bins, ec="k")
g.axes[-1].legend()
plt.show()
In [13]:
bins = np.linspace(df.age.min(), df.age.max(), 10)
g = sns.FacetGrid(df, col="Gender", hue="loan_status", palette="Set1", col_wrap=2)
g.map(plt.hist, 'age', bins=bins, ec="k")
g.axes[-1].legend()
plt.show()
Pre-processing: Feature selection/extraction
Lets look at the day of the week people get the loan
In [14]:
df['dayofweek'] = df['effective_date'].dt.dayofweek
bins = np.linspace(df.dayofweek.min(), df.dayofweek.max(), 10)
g = sns.FacetGrid(df, col="Gender", hue="loan_status", palette="Set1", col_wrap=2)
g.map(plt.hist, 'dayofweek', bins=bins, ec="k")
g.axes[-1].legend()
plt.show()
We see that people who get the loan at the end of the week dont pay it off, so lets use Feature binarization to set a threshold values less then day 4
In [15]:
df['weekend'] = df['dayofweek'].apply(lambda x: 1 if (x>3) else 0)
df.head()
Out[15]:
Unnamed: 0 Unnamed: 0.1 loan_status Principal terms effective_date due_date age education Gender dayofweek weekend
0 0 0 PAIDOFF 1000 30 2016-09-08 2016-10-07 45 High School or Below male 3 0
1 2 2 PAIDOFF 1000 30 2016-09-08 2016-10-07 33 Bechalor female 3 0
2 3 3 PAIDOFF 1000 15 2016-09-08 2016-09-22 27 college male 3 0
3 4 4 PAIDOFF 1000 30 2016-09-09 2016-10-08 28 college female 4 1
4 6 6 PAIDOFF 1000 30 2016-09-09 2016-10-08 29 college male 4 1
Convert Categorical features to numerical values
Lets look at gender:
In [16]:
df.groupby(['Gender'])['loan_status'].value_counts(normalize=True)
Out[16]:
Gender loan_status
female PAIDOFF 0.865385
COLLECTION 0.134615
male PAIDOFF 0.731293
COLLECTION 0.268707
Name: loan_status, dtype: float64
86 % of female pay there loans while only 73 % of males pay there loan
Lets convert male to 0 and female to 1:
In [17]:
df['Gender'].replace(to_replace=['male','female'], value=[0,1],inplace=True)
df.head()
Out[17]:
Unnamed: 0 Unnamed: 0.1 loan_status Principal terms effective_date due_date age education Gender dayofweek weekend
0 0 0 PAIDOFF 1000 30 2016-09-08 2016-10-07 45 High School or Below 0 3 0
1 2 2 PAIDOFF 1000 30 2016-09-08 2016-10-07 33 Bechalor 1 3 0
2 3 3 PAIDOFF 1000 15 2016-09-08 2016-09-22 27 college 0 3 0
3 4 4 PAIDOFF 1000 30 2016-09-09 2016-10-08 28 college 1 4 1
4 6 6 PAIDOFF 1000 30 2016-09-09 2016-10-08 29 college 0 4 1
One Hot Encoding
How about education?
In [18]:
df.groupby(['education'])['loan_status'].value_counts(normalize=True)
Out[18]:
education loan_status
Bechalor PAIDOFF 0.750000
COLLECTION 0.250000
High School or Below PAIDOFF 0.741722
COLLECTION 0.258278
Master or Above COLLECTION 0.500000
PAIDOFF 0.500000
college PAIDOFF 0.765101
COLLECTION 0.234899
Name: loan_status, dtype: float64
Feature befor One Hot Encoding
In [19]:
df[['Principal','terms','age','Gender','education']].head()
Out[19]:
Principal terms age Gender education
0 1000 30 45 0 High School or Below
1 1000 30 33 1 Bechalor
2 1000 15 27 0 college
3 1000 30 28 1 college
4 1000 30 29 0 college
Use one hot encoding technique to conver categorical varables to binary variables and append them to the feature Data Frame
In [20]:
Feature = df[['Principal','terms','age','Gender','weekend']]
Feature = pd.concat([Feature,pd.get_dummies(df['education'])], axis=1)
Feature.drop(['Master or Above'], axis = 1,inplace=True)
Feature.head()
Out[20]:
Principal terms age Gender weekend Bechalor High School or Below college
0 1000 30 45 0 0 0 1 0
1 1000 30 33 1 0 1 0 0
2 1000 15 27 0 0 0 0 1
3 1000 30 28 1 1 0 0 1
4 1000 30 29 0 1 0 0 1
Feature selection
Lets defind feature sets, X:
In [21]:
X = Feature
X[0:5]
Out[21]:
Principal terms age Gender weekend Bechalor High School or Below college
0 1000 30 45 0 0 0 1 0
1 1000 30 33 1 0 1 0 0
2 1000 15 27 0 0 0 0 1
3 1000 30 28 1 1 0 0 1
4 1000 30 29 0 1 0 0 1
What are our lables?
In [22]:
y = df['loan_status'].values
y[0:5]
Out[22]:
array(['PAIDOFF', 'PAIDOFF', 'PAIDOFF', 'PAIDOFF', 'PAIDOFF'],
dtype=object)
Normalize Data
Data Standardization give data zero mean and unit variance (technically should be done after train test split )
In [23]:
X= preprocessing.StandardScaler().fit(X).transform(X)
X[0:5]
Out[23]:
array([[ 0.51578458, 0.92071769, 2.33152555, -0.42056004, -1.20577805,
-0.38170062, 1.13639374, -0.86968108],
[ 0.51578458, 0.92071769, 0.34170148, 2.37778177, -1.20577805,
2.61985426, -0.87997669, -0.86968108],
[ 0.51578458, -0.95911111, -0.65321055, -0.42056004, -1.20577805,
-0.38170062, -0.87997669, 1.14984679],
[ 0.51578458, 0.92071769, -0.48739188, 2.37778177, 0.82934003,
-0.38170062, -0.87997669, 1.14984679],
[ 0.51578458, 0.92071769, -0.3215732, -0.42056004, 0.82934003,
-0.38170062, -0.87997669, 1.14984679]])
Classification
Now, it is your turn, use the training set to build an accurate model. Then use the test set to report the accuracy of the model You should use the following algorithm:
K Nearest Neighbor(KNN)
Decision Tree
Support Vector Machine
Logistic Regression
Notice:
You can go above and change the pre-processing, feature selection, feature-extraction, and so on, to make a better model.
You should use either scikit-learn, Scipy or Numpy libraries for developing the classification algorithms.
You should include the code of the algorithm in the following cells.
K Nearest Neighbor(KNN)
Notice: You should find the best k to build the model with the best accuracy.
warning: You should not use the loan_test.csv for finding the best k, however, you can split your train_loan.csv into train and test to find the best k.
In [24]:
#Train-Test Split
from sklearn.model_selection import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2)
In [25]:
#Training
from sklearn.neighbors import KNeighborsClassifier
from sklearn import metrics
Ks = 12
mean_acc = np.zeros((Ks-1))
std_acc = np.zeros((Ks-1))
ConfustionMtx=[];
for n in range(1,Ks):
neigh = KNeighborsClassifier(n_neighbors=n).fit(X_train, y_train)
yhat = neigh.predict(X_test)
mean_acc[n-1] = metrics.accuracy_score(y_test, yhat)
std_acc[n-1] = np.std(yhat==y_test)/np.sqrt(yhat.shape[0])
mean_acc
Out[25]:
array([0.71428571, 0.71428571, 0.71428571, 0.7 , 0.75714286,
0.71428571, 0.75714286, 0.72857143, 0.77142857, 0.72857143,
0.72857143])
In [26]:
plt.plot(range(1,Ks),mean_acc)
plt.fill_between(range(1,Ks),mean_acc - 1 * std_acc,mean_acc + 1 * std_acc, alpha=0.10)
plt.legend(('Accuracy ', '+/- 3xstd'))
plt.ylabel('Accuracy ')
plt.xlabel('Number of Neighbors (K)')
plt.tight_layout()
plt.show()
print( "The best accuracy was with", mean_acc.max(), "with k=", mean_acc.argmax()+1)
neigh = KNeighborsClassifier(n_neighbors=mean_acc.argmax()+1).fit(X_train, y_train)
The best accuracy was with 0.7714285714285715 with k= 9
In [90]:
print( "The best accuracy was with", mean_acc.max(), "with k=", mean_acc.argmax()+1)
The best accuracy was with 0.7857142857142857 with k= 1
In [91]:
# Set value of k as 7
k = 7
# Train Model and Predict
loanknn = KNeighborsClassifier(n_neighbors = k).fit(X_train,y_train)
loanknn
Out[91]:
KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',
metric_params=None, n_jobs=None, n_neighbors=7, p=2,
weights='uniform')
In [92]:
yhat = loanknn.predict(X_test)
yhat[0:5]
Out[92]:
array(['PAIDOFF', 'PAIDOFF', 'COLLECTION', 'COLLECTION', 'COLLECTION'],
dtype=object)
In [93]:
print("Train set Accuracy: ", metrics.accuracy_score(y_train, loanknn.predict(X_train)))
print("Test set Accuracy: ", metrics.accuracy_score(y_test, yhat))
Train set Accuracy: 0.8188405797101449
Test set Accuracy: 0.7222222222222222
In [94]:
from sklearn.metrics import classification_report
print (classification_report(y_test, yhat))
precision recall f1-score support
COLLECTION 0.44 0.29 0.35 14
PAIDOFF 0.78 0.88 0.82 40
accuracy 0.72 54
macro avg 0.61 0.58 0.59 54
weighted avg 0.69 0.72 0.70 54
In [95]:
from sklearn.metrics import f1_score
f1_score(y_test, yhat, average='weighted')
Out[95]:
0.7001989201477693
In [96]:
from sklearn.metrics import jaccard_similarity_score
jaccard_similarity_score(y_test, yhat)
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
Out[96]:
0.7222222222222222
In [ ]:
Decision Tree
In [97]:
# Import the decision tree model
from sklearn.tree import DecisionTreeClassifier
In [98]:
md = 10
mean_acc = np.zeros((md-1))
std_acc = np.zeros((md-1))
ConfustionMx = [];
for n in range(1,md):
#Train Model and Predict
loant = DecisionTreeClassifier(criterion="entropy", max_depth = n).fit(X_train,y_train)
yhat=loant.predict(X_test)
mean_acc[n-1] = metrics.accuracy_score(y_test, yhat)
std_acc[n-1]=np.std(yhat==y_test)/np.sqrt(yhat.shape[0])
mean_acc
Out[98]:
array([0.74074074, 0.74074074, 0.74074074, 0.75925926, 0.7962963 ,
0.77777778, 0.74074074, 0.72222222, 0.74074074])
In [99]:
plt.plot(range(1,md),mean_acc,'r')
plt.fill_between(range(1,md),mean_acc - 1 * std_acc,mean_acc + 1 * std_acc, alpha=0.10)
plt.legend(('Accuracy ', '+/- 3xstd'))
plt.ylabel('Accuracy ')
plt.xlabel('Number of Max Depth')
plt.tight_layout()
plt.show()
In [100]:
#Building the decision tree with max depth of 6
loandt = DecisionTreeClassifier(criterion="entropy", max_depth = 6)
# Check the default parameters
loandt
# Train the Decision tree model
loandt.fit(X_train,y_train)
# Predict using the model
yhat= loandt.predict(X_test)
In [101]:
#Calculating the train and test accuracy
print("Train set Accuracy: ", metrics.accuracy_score(y_train, loandt.predict(X_train)))
print("Test set Accuracy: ", metrics.accuracy_score(y_test, yhat))
#Building the confusion matrix
print (classification_report(y_test, yhat))
Train set Accuracy: 0.7934782608695652
Test set Accuracy: 0.7777777777777778
precision recall f1-score support
COLLECTION 0.57 0.57 0.57 14
PAIDOFF 0.85 0.85 0.85 40
accuracy 0.78 54
macro avg 0.71 0.71 0.71 54
weighted avg 0.78 0.78 0.78 54
In [102]:
# Calculate the F1 score
f1_score(y_test, yhat, average='weighted')
Out[102]:
0.7777777777777778
In [103]:
# Calculate the jaccard index
jaccard_similarity_score(y_test, yhat)
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
Out[103]:
0.7777777777777778
In [106]:
#Visualize the Decison tree
#!conda install -c conda-forge pydotplus -y
#!conda install -c conda-forge python-graphviz -y
In [107]:
'''from sklearn.externals.six import StringIO
import pydotplus
import matplotlib.image as mpimg
from sklearn import tree
%matplotlib inline '''
Out[107]:
'from sklearn.externals.six import StringIO\nimport pydotplus\nimport matplotlib.image as mpimg\nfrom sklearn import tree\n%matplotlib inline '
In [108]:
'''dot_data = StringIO()
filename = "loantree.png"
featureNames = Feature.columns
targetNames = df['loan_status'].unique().tolist()
out=tree.export_graphviz(loandt,feature_names=featureNames, out_file=dot_data, class_names= np.unique(y_train), filled=True, special_characters=True,rotate=False)
graph = pydotplus.graph_from_dot_data(dot_data.getvalue())
graph.write_png(filename)
img = mpimg.imread(filename)
plt.figure(figsize=(100, 200))
plt.imshow(img,interpolation='nearest')'''
Out[108]:
'dot_data = StringIO()\nfilename = "loantree.png"\nfeatureNames = Feature.columns\ntargetNames = df[\'loan_status\'].unique().tolist()\nout=tree.export_graphviz(loandt,feature_names=featureNames, out_file=dot_data, class_names= np.unique(y_train), filled=True, special_characters=True,rotate=False) \ngraph = pydotplus.graph_from_dot_data(dot_data.getvalue()) \ngraph.write_png(filename)\nimg = mpimg.imread(filename)\nplt.figure(figsize=(100, 200))\nplt.imshow(img,interpolation=\'nearest\')'
Support Vector Machine
In [109]:
# Import the library for SVM Classifier
from sklearn import svm
# Build a SVM Classifier with a Radial base Function Kernel
loansvm1 = svm.SVC(kernel='rbf').fit(X_train, y_train)
yhat1 = loansvm1.predict(X_test)
svm_r = metrics.accuracy_score(y_test, yhat1)
# Build a SVM Classifier with a Linear Kernel
loansvm2 = svm.SVC(kernel='linear').fit(X_train, y_train)
yhat2 = loansvm2.predict(X_test)
svm_l = metrics.accuracy_score(y_test, yhat2)
# Build a SVM Classifier with a Polynomial Kernel
loansvm3 = svm.SVC(kernel='poly').fit(X_train, y_train)
yhat3 = loansvm3.predict(X_test)
svm_p = metrics.accuracy_score(y_test, yhat3)
# Build a SVM Classifier with a Sigmoid Kernel
loansvm4 = svm.SVC(kernel='sigmoid').fit(X_train, y_train)
yhat4 = loansvm4.predict(X_test)
svm_s = metrics.accuracy_score(y_test, yhat4)
print(svm_r,svm_l,svm_p,svm_s)
0.7777777777777778 0.7407407407407407 0.7407407407407407 0.7037037037037037
In [110]:
# Find if labels are missing in the SVM models
print("The label missing in the first model with rbf kernel",set(y_test) - set(yhat1))
print("The label missing in the second model with linear",set(y_test) - set(yhat2))
print("The label missing in the third model with polynomial kernel",set(y_test) - set(yhat3))
print("The label missing in the fourth model with sigmoid kernel",set(y_test) - set(yhat4))
The label missing in the first model with rbf kernel set()
The label missing in the second model with linear {'COLLECTION'}
The label missing in the third model with polynomial kernel set()
The label missing in the fourth model with sigmoid kernel set()
In [111]:
#The SVM with the Radial base function and sigmoid kernel have the same accuracy (74.28%) and the models predicted the value collection as well, we conclude the that the rbf kernel is the best
#SVM Classifier with Radial base function kernel
# Build and train the SVM Classifier with a linear kernel
loansvm = svm.SVC(kernel='rbf').fit(X_train, y_train)
In [112]:
#Predicting the test values using the SVM model
yhat = loansvm.predict(X_test)
yhat [0:5]
Out[112]:
array(['PAIDOFF', 'PAIDOFF', 'PAIDOFF', 'PAIDOFF', 'COLLECTION'],
dtype=object)
In [113]:
print("Train set Accuracy: ", metrics.accuracy_score(y_train, loansvm.predict(X_train)))
print("Test set Accuracy: ", metrics.accuracy_score(y_test, yhat))
print (classification_report(y_test, yhat))
Train set Accuracy: 0.7681159420289855
Test set Accuracy: 0.7777777777777778
precision recall f1-score support
COLLECTION 0.67 0.29 0.40 14
PAIDOFF 0.79 0.95 0.86 40
accuracy 0.78 54
macro avg 0.73 0.62 0.63 54
weighted avg 0.76 0.78 0.74 54
In [114]:
# Calculate the f1 score
f1_score(y_test, yhat, average='weighted')
Out[114]:
0.7434343434343433
In [115]:
#Calculate the Jaccard index
jaccard_similarity_score(y_test, yhat)
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
Out[115]:
0.7777777777777778
Logistic Regression
In [116]:
# Import the library for Logistice regression
from sklearn.linear_model import LogisticRegression
# Build and train the logestic regression model
loanlr1 = LogisticRegression(C=0.01, solver='liblinear').fit(X_train,y_train)
yhat1 = loanlr1.predict(X_test)
loanlr_a1 = metrics.accuracy_score(y_test, yhat1)
# Build and train the logestic regression model
loanlr2 = LogisticRegression(C=0.01, solver='sag').fit(X_train,y_train)
yhat2 = loanlr2.predict(X_test)
loanlr_a2 = metrics.accuracy_score(y_test, yhat2)
# Build and train the logestic regression model
loanlr3 = LogisticRegression(C=0.01, solver='saga').fit(X_train,y_train)
yhat3 = loanlr3.predict(X_test)
loanlr_a3 = metrics.accuracy_score(y_test, yhat3)
# Build and train the logestic regression model
loanlr4 = LogisticRegression(C=0.01, solver='newton-cg').fit(X_train,y_train)
yhat4 = loanlr4.predict(X_test)
loanlr_a4 = metrics.accuracy_score(y_test, yhat4)
# Build and train the logestic regression model
loanlr5 = LogisticRegression(C=0.01, solver='lbfgs').fit(X_train,y_train)
yhat5 = loanlr5.predict(X_test)
loanlr_a5 = metrics.accuracy_score(y_test, yhat5)
print('LR model with liblinear solver',loanlr_a1)
print('LR model with sag solver',loanlr_a2)
print('LR model with saga solver',loanlr_a3)
print('LR model with newton-cg solver',loanlr_a4)
print('LR model with lbfgs solver',loanlr_a5)
LR model with liblinear solver 0.7592592592592593
LR model with sag solver 0.7407407407407407
LR model with saga solver 0.7407407407407407
LR model with newton-cg solver 0.7407407407407407
LR model with lbfgs solver 0.7407407407407407
In [117]:
# Find if labels are missing in the models
print("The label missing in the LR model with liblinear solver",set(y_test) - set(yhat1))
print("The label missing in the LR model with sag solver",set(y_test) - set(yhat2))
print("The label missing in the LR model with saga solver",set(y_test) - set(yhat3))
print("The label missing in the LR model with newton-cg solver",set(y_test) - set(yhat4))
print("The label missing in the LR model with lbfgs solver",set(y_test) - set(yhat5))
The label missing in the LR model with liblinear solver set()
The label missing in the LR model with sag solver {'COLLECTION'}
The label missing in the LR model with saga solver {'COLLECTION'}
The label missing in the LR model with newton-cg solver {'COLLECTION'}
The label missing in the LR model with lbfgs solver {'COLLECTION'}
In [118]:
#Except for the liblinear solver all other model has skipped the lable "collection" from the predicted values. Hence, the best logistic classifier will be the one with a liblinear solver
loanlr = LogisticRegression(C=0.01, solver='liblinear').fit(X_train,y_train)
yhat = loanlr.predict(X_test)
In [119]:
print("Train set Accuracy: ", metrics.accuracy_score(y_train, loanlr.predict(X_train)))
print("Test set Accuracy: ", metrics.accuracy_score(y_test, yhat))
print (classification_report(y_test, yhat))
Train set Accuracy: 0.7536231884057971
Test set Accuracy: 0.7592592592592593
precision recall f1-score support
COLLECTION 1.00 0.07 0.13 14
PAIDOFF 0.75 1.00 0.86 40
accuracy 0.76 54
macro avg 0.88 0.54 0.50 54
weighted avg 0.82 0.76 0.67 54
In [120]:
# Calculate the f1 score
f1_score(y_test, yhat, average='weighted')
Out[120]:
0.6717642373556352
In [121]:
#Calculate the Jaccard index
jaccard_similarity_score(y_test, yhat)
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
Out[121]:
0.7592592592592593
Model Evaluation using Test set
In [122]:
from sklearn.metrics import jaccard_similarity_score
from sklearn.metrics import f1_score
from sklearn.metrics import log_loss
First, download and load the test set:
Load Test set for evaluation
In [123]:
test_df = pd.read_csv('loan_test.csv')
test_df.head()
Out[123]:
Unnamed: 0 Unnamed: 0.1 loan_status Principal terms effective_date due_date age education Gender
0 1 1 PAIDOFF 1000 30 9/8/2016 10/7/2016 50 Bechalor female
1 5 5 PAIDOFF 300 7 9/9/2016 9/15/2016 35 Master or Above male
2 21 21 PAIDOFF 1000 30 9/10/2016 10/9/2016 43 High School or Below female
3 24 24 PAIDOFF 1000 30 9/10/2016 10/9/2016 26 college male
4 35 35 PAIDOFF 800 15 9/11/2016 9/25/2016 29 Bechalor male
In [124]:
# shape of the test data set
test_df.shape
Out[124]:
(54, 10)
In [125]:
# Count of the loan status
test_df['loan_status'].value_counts()
Out[125]:
PAIDOFF 40
COLLECTION 14
Name: loan_status, dtype: int64
In [126]:
df = test_df
df['due_date'] = pd.to_datetime(df['due_date'])
df['effective_date'] = pd.to_datetime(df['effective_date'])
df['dayofweek'] = df['effective_date'].dt.dayofweek
df['weekend'] = df['dayofweek'].apply(lambda x: 1 if (x>3) else 0)
df.groupby(['Gender'])['loan_status'].value_counts(normalize=True)
df['Gender'].replace(to_replace=['male','female'], value=[0,1],inplace=True)
df.groupby(['education'])['loan_status'].value_counts(normalize=True)
Feature = df[['Principal','terms','age','Gender','weekend']]
Feature = pd.concat([Feature,pd.get_dummies(df['education'])], axis=1)
Feature.drop(['Master or Above'], axis = 1,inplace=True)
X_test = Feature
y_test = df['loan_status'].values
X_test = preprocessing.StandardScaler().fit(X_test).transform(X_test)
In [128]:
# KNN model testing
yhat_knn = loanknn.predict(X_test)
# Calculate the f1 score
f1_knn = f1_score(y_test, yhat_knn, average='weighted')
#Calculate the Jaccard index# Predict using the model
jsc_knn = jaccard_similarity_score(y_test, yhat_knn)
print('f1 score: ',f1_knn)
print('Jaccard index: ',jsc_knn)
f1 score: 0.7001989201477693
Jaccard index: 0.7222222222222222
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
In [132]:
# Predict using the model
yhat_dt= loandt.predict(X_test)
# Calculate the f1 score
f1_dt = f1_score(y_test, yhat_dt, average='weighted')
#Calculate the Jaccard index# Predict using the model
jsc_dt = jaccard_similarity_score(y_test, yhat_dt)
print('f1 score: ',f1_dt)
print('Jaccard index: ',jsc_dt)
f1 score: 0.7777777777777778
Jaccard index: 0.7777777777777778
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
In [133]:
# Predict using the model
yhat_svm = loansvm.predict(X_test)
# Calculate the f1 score
f1_svm = f1_score(y_test, yhat_svm, average='weighted')
#Calculate the Jaccard index# Predict using the model
jsc_svm = jaccard_similarity_score(y_test, yhat_svm)
print('f1 score: ',f1_svm)
print('Jaccard index: ',jsc_svm)
f1 score: 0.7434343434343433
Jaccard index: 0.7777777777777778
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
In [134]:
# Predict using the model
yhat_lr = loanlr.predict(X_test)
# Calculate the f1 score
f1_lr = f1_score(y_test, yhat_lr, average='weighted')
#Calculate the Jaccard index# Predict using the model
jsc_lr = jaccard_similarity_score(y_test, yhat_lr)
# Calculate Log loss
yhat_lr_prob = loanlr.predict_proba(X_test)
ll_lr = log_loss(y_test, yhat_lr_prob)
print('f1 score: ',f1_lr)
print('Jaccard index: ',jsc_lr)
print('Log Loss: ',ll_lr)
f1 score: 0.6717642373556352
Jaccard index: 0.7592592592592593
Log Loss: 0.5693569109817576
A:\Anaconda3\lib\site-packages\sklearn\metrics\_classification.py:664: FutureWarning: jaccard_similarity_score has been deprecated and replaced with jaccard_score. It will be removed in version 0.23. This implementation has surprising behavior for binary and multiclass classification tasks.
FutureWarning)
In [135]:
Jaccard = [jsc_knn,jsc_dt,jsc_svm,jsc_lr]
F1_score = [f1_knn,f1_dt,f1_svm,f1_lr]
LogLoss = ['NA','NA','NA',ll_lr]
df = {'Algorithm': ['KNN', 'Decistion Tree', 'SVM', 'LogisticRegression'], \
'Jaccard': Jaccard, 'F1-score': F1_score, 'LogLoss': LogLoss}
Report = pd.DataFrame(data=df, columns=['Algorithm', 'Jaccard', 'F1-score', 'LogLoss'], index=None)
Report
Out[135]:
Algorithm Jaccard F1-score LogLoss
0 KNN 0.722222 0.700199 NA
1 Decistion Tree 0.777778 0.777778 NA
2 SVM 0.777778 0.743434 NA
3 LogisticRegression 0.759259 0.671764 0.569357
Report
You should be able to report the accuracy of the built model using different evaluation metrics:
Algorithm Jaccard F1-score LogLoss
KNN ? ? NA
Decision Tree ? ? NA
SVM ? ? NA
LogisticRegression ? ? ?
Want to learn more?
IBM SPSS Modeler is a comprehensive analytics platform that has many machine learning algorithms. It has been designed to bring predictive intelligence to decisions made by individuals, by groups, by systems – by your enterprise as a whole. A free trial is available through this course, available here: SPSS Modeler
Also, you can use Watson Studio to run these notebooks faster with bigger datasets. Watson Studio is IBM's leading cloud solution for data scientists, built by data scientists. With Jupyter notebooks, RStudio, Apache Spark and popular libraries pre-packaged in the cloud, Watson Studio enables data scientists to collaborate on their projects without having to install anything. Join the fast-growing community of Watson Studio users today with a free account at Watson Studio
Thanks for completing this lesson!
Author: Saeed Aghabozorgi
Saeed Aghabozorgi, PhD is a Data Scientist in IBM with a track record of developing enterprise level applications that substantially increases clients’ ability to turn data into actionable knowledge. He is a researcher in data mining field and expert in developing advanced analytic methods like machine learning and statistical modelling on large datasets.
Copyright © 2018 Cognitive Class. This notebook and its source code are released under the terms of the MIT License.-
This topic was modified 2 years, 9 months ago by
Abhishek Tyagi.
