done

Author: deruhat

Diff for: README.md

@@ -0,0 +1 @@
+# Image-Classification

Diff for: answers.py

@@ -0,0 +1,13 @@
+# answers.py
+# ----------
+# Licensing Information: Please do not distribute or publish solutions to this
+# project. You are free to use and extend these projects for educational
+# purposes. The Pacman AI projects were developed at UC Berkeley, primarily by
+# John DeNero ([email protected]) and Dan Klein ([email protected]).
+# For more info, see http://inst.eecs.berkeley.edu/~cs188/sp09/pacman.html
+
+def q2():
+  "*** YOUR CODE HERE ***"
+
+def q4():
+  "*** YOUR CODE HERE ***"

Diff for: classification.html

@@ -0,0 +1,57 @@
+# classificationMethod.py
+# -----------------------
+# Licensing Information: Please do not distribute or publish solutions to this
+# project. You are free to use and extend these projects for educational
+# purposes. The Pacman AI projects were developed at UC Berkeley, primarily by
+# John DeNero ([email protected]) and Dan Klein ([email protected]).
+# For more info, see http://inst.eecs.berkeley.edu/~cs188/sp09/pacman.html
+
+# This file contains the abstract class ClassificationMethod
+
+class ClassificationMethod:
+  """
+  ClassificationMethod is the abstract superclass of 
+   - MostFrequentClassifier
+   - NaiveBayesClassifier
+   - PerceptronClassifier
+   - MiraClassifier
+ 
+  As such, you need not add any code to this file.  You can write
+  all of your implementation code in the files for the individual
+  classification methods listed above.
+  """
+  def __init__(self, legalLabels):
+    """
+    For digits dataset, the set of legal labels will be 0,1,..,9
+    For faces dataset, the set of legal labels will be 0 (non-face) or 1 (face)
+    """
+    self.legalLabels = legalLabels
+    
+    
+  def train(self, trainingData, trainingLabels, validationData, validationLabels):
+    """
+    This is the supervised training function for the classifier.  Two sets of 
+    labeled data are passed in: a large training set and a small validation set.
+    
+    Many types of classifiers have a common training structure in practice: using
+    training data for the main supervised training loop but tuning certain parameters
+    with a small held-out validation set.
+
+    For some classifiers (naive Bayes, MIRA), you will need to return the parameters' 
+    values after traning and tuning step.
+    
+    To make the classifier generic to multiple problems, the data should be represented
+    as lists of Counters containing feature descriptions and their counts.
+    """
+    abstract
+    
+  def classify(self, data):
+    """
+    This function returns a list of labels, each drawn from the set of legal labels
+    provided to the classifier upon construction.
+
+    To make the classifier generic to multiple problems, the data should be represented
+    as lists of Counters containing feature descriptions and their counts.
+    """
+    abstract
+

Diff for: classificationMethod.py

@@ -0,0 +1,10 @@
+python dataClassifier.py  
+python dataClassifier.py -h  
+python dataClassifier.py -c naiveBayes --autotune 
+python dataClassifier.py -a -d digits -c naiveBayes -o -1 3 -2 6  
+python dataClassifier.py -c perceptron 
+python dataClassifier.py -c perceptron -w  
+python dataClassifier.py -c mira --autotune 
+python dataClassifier.py -d digits -c naiveBayes -f -a -t 1000  
+python dataClassifier.py -d digits -c minicontest
+python runMinicontest.py