Welcome to the second project of the Machine Learning Engineer Nanodegree! In this notebook, some template code has already been provided for you, and it will be your job to implement the additional functionality necessary to successfully complete this project. Sections that begin with 'Implementation' in the header indicate that the following block of code will require additional functionality which you must provide. Instructions will be provided for each section and the specifics of the implementation are marked in the code block with a 'TODO'
statement. Please be sure to read the instructions carefully!
In addition to implementing code, there will be questions that you must answer which relate to the project and your implementation. Each section where you will answer a question is preceded by a 'Question X' header. Carefully read each question and provide thorough answers in the following text boxes that begin with 'Answer:'. Your project submission will be evaluated based on your answers to each of the questions and the implementation you provide.
Note: Code and Markdown cells can be executed using the Shift + Enter keyboard shortcut. In addition, Markdown cells can be edited by typically double-clicking the cell to enter edit mode.
Answer:
The type of supervised learning problem is derived from the kind of output we are expecting. In regression we match inputs to continuous outputs, or in other words we predict the actual numeric value a specific input will have. On the other hand, in classification inputs are mapped into discrete outputs (Boolean values for example).
For this specific problem we need to detect those students who are projected to fail and need intervention. In order to reach this goal, the final dataset will be divided into two sets: Those who need intervention and does who do not. Making the problem a boolean variable and hence a classification problem
In [9]:
# Import libraries
import numpy as np
import pandas as pd
from time import time
from sklearn.metrics import f1_score
# Read student data
student_data = pd.read_csv("student-data.csv")
print "Student data read successfully!"
Let's begin by investigating the dataset to determine how many students we have information on, and learn about the graduation rate among these students. In the code cell below, you will need to compute the following:
n_students
.n_features
.n_passed
.n_failed
.grad_rate
, in percent (%).
In [10]:
# TODO: Calculate number of students
n_students = len(student_data)
# TODO: Calculate number of features
n_features = len(student_data.columns[:-1])
# TODO: Calculate passing students
n_passed = len(student_data[student_data['passed'] == 'yes'])
# TODO: Calculate failing students
n_failed = n_students - n_passed
# TODO: Calculate graduation rate
grad_rate = (float(n_passed) / n_students) * 100
# Print the results
print "Total number of students: {}".format(n_students)
print "Number of features: {}".format(n_features)
print "Number of students who passed: {}".format(n_passed)
print "Number of students who failed: {}".format(n_failed)
print "Graduation rate of the class: {:.2f}%".format(grad_rate)
In this section, we will prepare the data for modeling, training and testing.
It is often the case that the data you obtain contains non-numeric features. This can be a problem, as most machine learning algorithms expect numeric data to perform computations with.
Run the code cell below to separate the student data into feature and target columns to see if any features are non-numeric.
In [11]:
# Extract feature columns
feature_cols = list(student_data.columns[:-1])
# Extract target column 'passed'
target_col = student_data.columns[-1]
# Show the list of columns
print "Feature columns:\n{}".format(feature_cols)
print "\nTarget column: {}".format(target_col)
# Separate the data into feature data and target data (X_all and y_all, respectively)
X_all = student_data[feature_cols]
y_all = student_data[target_col]
from IPython.display import display
# Show the feature information by printing the first five rows
print "\nFeature values:"
# Added pretty table display
display(X_all.head())
As you can see, there are several non-numeric columns that need to be converted! Many of them are simply yes
/no
, e.g. internet
. These can be reasonably converted into 1
/0
(binary) values.
Other columns, like Mjob
and Fjob
, have more than two values, and are known as categorical variables. The recommended way to handle such a column is to create as many columns as possible values (e.g. Fjob_teacher
, Fjob_other
, Fjob_services
, etc.), and assign a 1
to one of them and 0
to all others.
These generated columns are sometimes called dummy variables, and we will use the pandas.get_dummies()
function to perform this transformation. Run the code cell below to perform the preprocessing routine discussed in this section.
In [12]:
def preprocess_features(X):
''' Preprocesses the student data and converts non-numeric binary variables into
binary (0/1) variables. Converts categorical variables into dummy variables. '''
# Initialize new output DataFrame
output = pd.DataFrame(index = X.index)
# Investigate each feature column for the data
for col, col_data in X.iteritems():
# If data type is non-numeric, replace all yes/no values with 1/0
if col_data.dtype == object:
col_data = col_data.replace(['yes', 'no'], [1, 0])
# If data type is categorical, convert to dummy variables
if col_data.dtype == object:
# Example: 'school' => 'school_GP' and 'school_MS'
col_data = pd.get_dummies(col_data, prefix = col)
# Collect the revised columns
output = output.join(col_data)
return output
X_all = preprocess_features(X_all)
print "Processed feature columns ({} total features):\n{}".format(len(X_all.columns), list(X_all.columns))
In [13]:
%matplotlib inline
import seaborn as sns
sns.factorplot("failures", col="goout", data=student_data, hue='passed', kind="count");
So far, we have converted all categorical features into numeric values. For the next step, we split the data (both features and corresponding labels) into training and test sets. In the following code cell below, you will need to implement the following:
X_all
, y_all
) into training and testing subsets.random_state
for the function(s) you use, if provided.X_train
, X_test
, y_train
, and y_test
.
In [14]:
# TODO: Import any additional functionality you may need here
# TODO: Set the number of training points
num_train = 300
# Set the number of testing points
num_test = X_all.shape[0] - num_train
# TODO: Shuffle and split the dataset into the number of training and testing points above
from sklearn.cross_validation import train_test_split
X_train, X_test, y_train, y_test = train_test_split(X_all, y_all, stratify=y_all,train_size=num_train, random_state=37)
# Show the results of the split
print "Training set has {} samples.".format(X_train.shape[0])
print "Testing set has {} samples.".format(X_test.shape[0])
print "\nTrain set 'yes' pct = {:.2f}%".format(100 * (y_train == 'yes').mean())
print "Test set 'yes' pct = {:.2f}%".format(100 * (y_test == 'yes').mean())
In this section, you will choose 3 supervised learning models that are appropriate for this problem and available in scikit-learn
. You will first discuss the reasoning behind choosing these three models by considering what you know about the data and each model's strengths and weaknesses. You will then fit the model to varying sizes of training data (100 data points, 200 data points, and 300 data points) and measure the F1 score. You will need to produce three tables (one for each model) that shows the training set size, training time, prediction time, F1 score on the training set, and F1 score on the testing set.
The following supervised learning models are currently available in scikit-learn
that you may choose from:
List three supervised learning models that are appropriate for this problem. For each model chosen
Answer:
Run the code cell below to initialize three helper functions which you can use for training and testing the three supervised learning models you've chosen above. The functions are as follows:
train_classifier
- takes as input a classifier and training data and fits the classifier to the data.predict_labels
- takes as input a fit classifier, features, and a target labeling and makes predictions using the F1 score.train_predict
- takes as input a classifier, and the training and testing data, and performs train_clasifier
and predict_labels
.
In [15]:
def train_classifier(clf, X_train, y_train):
''' Fits a classifier to the training data. '''
# Start the clock, train the classifier, then stop the clock
start = time()
clf.fit(X_train, y_train)
end = time()
# Print the results
print "Trained model in {:.4f} seconds".format(end - start)
def predict_labels(clf, features, target):
''' Makes predictions using a fit classifier based on F1 score. '''
# Start the clock, make predictions, then stop the clock
start = time()
y_pred = clf.predict(features)
end = time()
# Print and return results
print "Made predictions in {:.4f} seconds.".format(end - start)
return f1_score(target.values, y_pred, pos_label='yes')
def train_predict(clf, X_train, y_train, X_test, y_test):
''' Train and predict using a classifer based on F1 score. '''
# Indicate the classifier and the training set size
print "Training a {} using a training set size of {}. . .".format(clf.__class__.__name__, len(X_train))
# Train the classifier
train_classifier(clf, X_train, y_train)
f1_training = predict_labels(clf, X_train, y_train)
f1_test = predict_labels(clf, X_test, y_test)
# Print the results of prediction for both training and testing
print "F1 score for training set: {:.4f}.".format(f1_training)
print "F1 score for test set: {:.4f}.".format(f1_test)
return [f1_training, f1_test]
With the predefined functions above, you will now import the three supervised learning models of your choice and run the train_predict
function for each one. Remember that you will need to train and predict on each classifier for three different training set sizes: 100, 200, and 300. Hence, you should expect to have 9 different outputs below — 3 for each model using the varying training set sizes. In the following code cell, you will need to implement the following:
clf_A
, clf_B
, and clf_C
.random_state
for each model you use, if provided.X_train
and y_train
.
In [16]:
# TODO: Import the three supervised learning models from sklearn
from sklearn import tree
from sklearn import svm
from sklearn.ensemble import GradientBoostingClassifier
#from sklearn.ensemble import RandomForestClassifier
# TODO: Initialize the three models
rand_state = 37
clf_A = tree.DecisionTreeClassifier(random_state=rand_state)
clf_B = svm.SVC(random_state=rand_state)
clf_C = GradientBoostingClassifier(random_state=rand_state)
models = [clf_A, clf_B, clf_C]
# TODO: Execute the 'train_predict' function for each classifier and each training set size
# train_predict(clf, X_train, y_train, X_test, y_test)
results = []
for model in models:
# TODO: Set up the training set sizes
print '************************************************************\n'
train = []
test = []
for set_size in (100,200,300):
print '------------------------------------------------------------'
train_values = train_predict(model, X_train[:set_size], y_train[:set_size], X_test, y_test)
train.append(train_values[0])
test.append(train_values[1])
results.append(train)
results.append(test)
Edit the cell below to see how a table can be designed in Markdown. You can record your results from above in the tables provided.
Classifer 1 - Decision Tree Classifier
Training Set Size | Training Time | Prediction Time (test) | F1 Score (train) | F1 Score (test) |
---|---|---|---|---|
100 | 0.0012 | 0.0004 | 1.00 | 0.6552 |
200 | 0.0015 | 0.0003 | 1.00 | 0.7231 |
300 | 0.0022 | 0.0003 | 1.00 | 0.7244 |
Classifer 2 - Support Vector Machines (SVM)
Training Set Size | Training Time | Prediction Time (test) | F1 Score (train) | F1 Score (test) |
---|---|---|---|---|
100 | 0.0010 | 0.009 | 0.8444 | 0.7287 |
200 | 0.0042 | 0.0016 | 0.8525 | 0.7755 |
300 | 0.0085 | 0.0021 | 0.8720 | 0.8212 |
Classifer 3 - Ensemble Gradient Boosting
Training Set Size | Training Time | Prediction Time (test) | F1 Score (train) | F1 Score (test) |
---|---|---|---|---|
100 | 0.0647 | 0.0005 | 1.0000 | 0.6723 |
200 | 0.0808 | 0.0004 | 0.9962 | 0.7669 |
300 | 0.0933 | 0.0004 | 0.9756 | 0.7536 |
In [17]:
import matplotlib.pyplot as plt
%matplotlib inline
for a in range(3):
plt.plot([100,200,300], results[a+a], '-o')
plt.legend(['y = Tree_train', 'y = SVM_train', 'y = EGB_train'],
bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0)
plt.ylabel('Score')
plt.title('Training Scores')
plt.show()
for a in range(1,6,2):
plt.plot([100,200,300], results[a], '-o')
plt.legend(['y = Tree_test', 'y = SVM_test', 'y = EGB_test'],
bbox_to_anchor=(1.05, 1), loc=2, borderaxespad=0)
plt.ylabel('Score')
plt.title('Testing Scores')
plt.show()
In this final section, you will choose from the three supervised learning models the best model to use on the student data. You will then perform a grid search optimization for the model over the entire training set (X_train
and y_train
) by tuning at least one parameter to improve upon the untuned model's F1 score.
Based on the experiments you performed earlier, in one to two paragraphs, explain to the board of supervisors what single model you chose as the best model. Which model is generally the most appropriate based on the available data, limited resources, cost, and performance?
Answer:
Dear Board of supervisors,
Given all the previous results we will be implementing the Support Vector Machine model to predict which students are in need of tutoring in order to prevent them from failing. In terms of performance based on time, decision trees gives us the faster options in terms of training and testing time. However, having a perfect score on the training is not ideal in the case of the Decision tree, cause is it an indication of overfitting. This is shown in the accuracy results on the testing set by obtaining the worst outcome of the three algorithms. In comparison with Gradient Boosting, it is slower in terms of prediction but greatly faster in terms of the training set. Gradient Boosting getting close to a perfect score in the testing set is not an indication of overfitting due to the sequential building of trees the algorithm uses, which also makes a greater impact in terms of the time spent training the data.
In conclusion, SVM provides to be the best option in terms of quality since it provides the best possible score, while also being timely efficient in the training and prediction aspects. It is also the algorithm that provides better tuning options by the ability of using a different kernel in order to improve performance. I am confident at the end of the tuning a prediction of more than 80% accuracy will be warranted.
In one to two paragraphs, explain to the board of directors in layman's terms how the final model chosen is supposed to work. Be sure that you are describing the major qualities of the model, such as how the model is trained and how the model makes a prediction. Avoid using advanced mathematical or technical jargon, such as describing equations or discussing the algorithm implementation.
Answer: Dear Board of Directors,
We are currently facing a classification problem, which means we need to use all our data to determine if one element belongs to it or not based on the attributes. For example, when determining if a vehicle is a car or a motorcycle we count the number of wheels; if 2 wheels are present then is a motorcyle else it is a car. Our goal is to find out if a student will fail or not the class based on the given parameters.
This is precisely what a Support Vector Machine does. In order to explain it let us think about a football game. During the game we have 22 players on the field, for which 11 belongs to Ohio State and 11 belong to Michigan. Before each play starts there is an imaginary line that divides the start point of those two teams known as the line of scrimmage. What SVM does is try to find this line of scrimmage in our data, so in the case another player is added to the Michigan side, we can surely predict is in fact a Michigan player (and an illegal formation penalty for them of course).
In SVM we might have some ignorant players rooming in the wrong side of the field, so the line of scrimmage would not precisely give us a clean separation. But it will clearly give us a line that maximizes the division between this two categories. It also inclues a tric play, known as kernel trick, that allow us to add new dimensions to our data in order to make better line separations.
Fine tune the chosen model. Use grid search (GridSearchCV
) with at least one important parameter tuned with at least 3 different values. You will need to use the entire training set for this. In the code cell below, you will need to implement the following:
sklearn.grid_search.GridSearchCV
and sklearn.metrics.make_scorer
.parameters = {'parameter' : [list of values]}
.clf
.make_scorer
and store it in f1_scorer
.pos_label
parameter to the correct value!clf
using f1_scorer
as the scoring method, and store it in grid_obj
.X_train
, y_train
), and store it in grid_obj
.
In [25]:
# TODO: Import 'GridSearchCV' and 'make_scorer'
from sklearn.metrics import make_scorer
from sklearn.grid_search import GridSearchCV
# TODO: Create the parameters list you wish to tune
parameters = {'kernel':['linear', 'rbf', 'poly','sigmoid'],
'C': [0.6, 1, 1.5, 3],
'probability': [True, False],
'tol': [1e-6,1e-5, 1e-4],
'random_state': [37]
}
# TODO: Initialize the classifier
clf = svm.SVC()
# TODO: Make an f1 scoring function using 'make_scorer'
def f1_metrics(y_true, y_pred):
f1 = f1_score(y_true, y_pred, pos_label='yes')
return f1
f1_scorer = make_scorer(f1_metrics)
# TODO: Perform grid search on the classifier using the f1_scorer as the scoring method
grid_obj = GridSearchCV(clf, param_grid=parameters, scoring=f1_scorer)
# TODO: Fit the grid search object to the training data and find the optimal parameters
grid_obj = grid_obj.fit(X_train, y_train)
# Get the estimator
clf = grid_obj.best_estimator_
# Report the final F1 score for training and testing after parameter tuning
print clf
print "Tuned model has a training F1 score of {:.4f}.".format(predict_labels(clf, X_train, y_train))
print "Tuned model has a testing F1 score of {:.4f}.".format(predict_labels(clf, X_test, y_test))
In [61]:
from IPython.display import display
#display(pd.DataFrame(grid_obj.grid_scores_))
grid_results = pd.DataFrame(grid_obj.grid_scores_)
grid_ = [[x[0]['C'], x[0]['tol'], x[1]] for index, x in list(grid_results.iterrows())]
#sns.heatmap( grid_, annot=True, fmt="d", linewidths=.5)
# tried but couldn't make it a heatmap
Answer:
Model | Training Time | Prediction Time (test) | F1 Score (train) | F1 Score (test) |
---|---|---|---|---|
Un-tuned | 0.0085 | 0.0021 | 0.8650 | 0.8212 |
Tuned | 0.0056 | 0.0020 | 0.8358 | 0.8205 |
Overall by tuning the model we were able to siglithly increase the testing scores in the model but made a huge imporvement in terms of time efficiency.
In [19]:
students = X_train[:10]
display(students)
Prediction for them:
In [20]:
for i, price in enumerate(clf.predict(students)):
print "Is the Student {} predicted to pass the year: {}".format(i+1, price)
In [21]:
X_all.iloc[164]
Out[21]:
Based on this induvidual study, we are confident our model correctly predicted his outcome as failing.
Note: Once you have completed all of the code implementations and successfully answered each question above, you may finalize your work by exporting the iPython Notebook as an HTML document. You can do this by using the menu above and navigating to
File -> Download as -> HTML (.html). Include the finished document along with this notebook as your submission.