Provide accurate and complete reports Sub Total & comments Provide relatively accurate and complete reports III. Provide inaccurate and/or incomplete reports III. Perform all required tasks correctly and consistently 20 III. Perform most of the required tasks correctly and consistently II. Do not perform all required tasks correctly and consistently II. Identify and use correct use correct algorithm for algorithm for the problem in the problem in hand hand II. Do not use correct algorithm(s) for the problem in hand (50%-75%) (76%-100%) IV. In your report include the snapshot of the decision tree, confusion matrix and classification accuracy and other descriptionsĬBOK Satisfactory Good Unsatisfactory (0%-49%) ) Marks Allocated % Marks Attained Data and Information Management IV. Finally using the best combination of parameters and features, construct the decision tree. Explain your findings in relation to the meaning of the parameters. Also report the size of the trees in a separate table. C is the confidence factor used for pruning, M is the minimum instance per leaf Report the error rate on the training set of trees built with these different parameter settings. The classification algorithm on Weka has two important parameters, C and M. Fill in the table below Feature(s)eliminated Resulting accuracy) 6. Experiment with the different attribute evaluation methods such as Chi Square, Information Gain, and Gain ratio. Using the visualization and attribute selection menu, identify the features, which are least important for classification and therefore can be eliminated. We will be using 10 fold cross validation for all the steps here on.
Why is the cross validation a better choice for testing? 5. What is the resulting classification accuracy? 4. Now set the test option to 10 fold cross-validation. Set the test option of the classifier to "Use training set" What is the classification accuracy? 3. Construct the 148 decision tree using all the attributes. Import the tae.arff dataset into WEKA explorer.
0 kommentar(er)
