Programming Project 2 Along Party Lines This problem set is due 11 20 Since it is an important year for U S elections we ll do a political programming project In a separate file we ve included the party membership and voting records of various member of congress We d like you to use this data to classify congressmen as Republicans or Democrats We d like you to use Bayesian Classifiers and Decision Trees We ve included their votes on a number of issues The data set is from the 80s so the issues aren t current you can envision the congressmen with funny hair if you d like And the issues are 1 Handicapped Infants 2 Water Project Cost Sharing 3 Adopt the Budget Resolution 4 Physician Fee Freeze 5 Aid to El Salvador 6 Religious Groups in Schools 7 Anti Satellite Test Ban 8 Aid to Nicaraguan Contras 9 MX Missile 10 Immigration 11 Synfuels Corporation Cutback 12 Education Spending 13 Superfund Right to Sue 14 Crime 15 Duty Free Exports 16 South Africa Export Administration Act 1 An individual congressmen is represented by a line in the training file no yes yes no no maybe yes yes yes yes yes no maybe yes yes yes democrat Which shows their votes on the issues and their political party Congressmen who didn t vote during a particular vote are given a maybe You may notice that some congressmen have not voted at all You have three files 1 training large csv The main training set 2 training small csv A smaller training set 3 testing csv The testing set 1 Part 1 Classification For each part create a classifier using the suggested python modules Train the classifier on the small training set and then test using the test data Collect precision and recall information Do the same thing for the larger training set Please document your code and make the structure clean readable and intuitive 1 1 Part 1a Bayesian Classifier Use a na ve Bayes classifier to classify the politicians You should use the Reverend classifier http divmod org trac wiki DivmodReverend You cannot simply just plug in the classifier you will have to write a tokenizer to split up the training input appropriately 1 2 Part 1b Decision Trees You can find a Python implimentation of the ID3 algorithm here along with a long tutorial http www onlamp com pub a python 2006 02 09 ai decision trees html 2 Part 2 Discussion Now that you ve classified things please discuss the different classification techniques Did different classifiers do better with this data set How did the size of the training set effect the results Did different methods work better with different size training sets This should definitely be longer than a paragraph and should be as long as you need to get your ideas across 2