diff --git a/C-1.png b/C-1.png
new file mode 100644
index 0000000..87a144f
Binary files /dev/null and b/C-1.png differ
diff --git a/C-50.png b/C-50.png
new file mode 100644
index 0000000..42e5157
Binary files /dev/null and b/C-50.png differ
diff --git a/learning_curve.py b/learning_curve.py
index 2364f2c..fdd6d6a 100644
--- a/learning_curve.py
+++ b/learning_curve.py
@@ -6,9 +6,20 @@
 from sklearn.cross_validation import train_test_split
 from sklearn.linear_model import LogisticRegression
 
+""" recognize images of handwritten digits """
+# digits = load_digits()
+# print digits.DESCR
+# fig = plt.figure()
+# for i in range(10):
+#     subplot = fig.add_subplot(5,2,i+1)
+#     subplot.matshow(numpy.reshape(digits.data[i],(8,8)),cmap='gray')
+
+# plt.show()
+
+""" initial conditions """
 data = load_digits()
-print data.DESCR
-num_trials = 10
+# print data.DESCR
+num_trials = 100
 train_percentages = range(5,95,5)
 test_accuracies = numpy.zeros(len(train_percentages))
 
@@ -17,10 +28,22 @@
 # You should repeat each training percentage num_trials times to smooth out variability
 # for consistency with the previous example use model = LogisticRegression(C=10**-10) for your learner
 
-# TODO: your code here
+""" partition data into two sets--training set and testing set 
+	vary training set size vs. testing set size and plot resulting curve """
+for i in range(len(train_percentages)):
+    trial_accuracies= []
+    for j in range(num_trials):
+        x_train,x_test,y_train,y_test = train_test_split(data.data, data.target, train_size=train_percentages[i]/200.0)                                                      
+        model = LogisticRegression(C=10 ** -10)
+        model.fit(x_train, y_train)
+        accur_score = model.score(x_test,y_test)
+        trial_accuracies.append(accur_score)
+
+    test_accuracies[i] = sum(trial_accuracies) / num_trials
 
 fig = plt.figure()
 plt.plot(train_percentages, test_accuracies)
 plt.xlabel('Percentage of Data Used for Training')
 plt.ylabel('Accuracy on Test Set')
 plt.show()
+
diff --git a/questions.txt b/questions.txt
new file mode 100644
index 0000000..471731b
--- /dev/null
+++ b/questions.txt
@@ -0,0 +1,9 @@
+1. What is the general trend in the curve?
+As percentage of data used for training increases, so does the accuracy on the test set.
+2. Are there parts of the curve that appear to be noisier than others?  Why?
+The curve between 35%/40% to 60% is generally noisier than the others. This is probably because when the test size is super small, there is a high chance that the accuracy will be low, just as when the test size is super large, there is a high chance that the accuracy will be high. However, in the middle, there is more uncertainty since the test size cannot determine the accuracy as well (because there are not enough trials in this range), and so the graph demonstrates this uncertainty by being noisier.
+3. How many trials do you need to get a smooth curve?
+The curve gets pretty smooth by 500 trials, but gets very smooth by 800 trials.
+4. Try different values for C (by changingLogisticRegression(C=10**-10)).  What happens?  
+When C is larger (ex: C = (10**-1), the curve more closely resembles a logarithmic curve; it's a smoother curve; it reaches a higher accuracy (0.96). When C is smaller (ex: C = (10**-50), the graph doesn't resemble any type of curve; it's noisier all around; it peaks to a smaller and smaller accuracy (0.098).
+