Data Mining and Machine Learning

Question: Discuss about theData Mining and Machine Learning. Answer: Preprocessing in Data Mining Steps for Preprocessing Data preprocessing method have four methods. They are, Data Cleaning Data Discretization Data Transformation Data Reduction Conclusion of Data Pre-Processing Steps In preprocessing stage, the training data of the file is modified to understandable format. The user viewed the data and understand the attributes available on the training data set. User used the filter option and views the data as clean. Then the user did group the data as based on the data type. The numerical data and the character data formed as group separately. The decimal value is changed as a whole value by the transformation method. Then the data reduction method is done in preprocessing stage. Finally got fully preprocessed data. The data is a clean data. Decision Tree explanation and why it has been chosen Decision Tree is a class for generating a pruned or unpruned C4.5 decision tree. Here in decision tree target class attribute as Year_Of_Study here goes to leaf and other attributes to branches. J28 Decision tree is used to classify the new item of the given data. For create the new item, first need to crate the decision tree. This decision tree made by using the attributes of the given training data. Particularly, the J48 decision tree is notify the attributes that clearly view the various instances when the training set is encounters. This is main advantage of the J48 algorithm. It gives the detail information about the all data instance. So, the user get the highest information of the training set data. In all the possible values of data, user take the target value as which data value is not provide the ambiguity. No ambiguity of data together to form a category. This category have the same target value. After obtain the target value user will terminate the branch. Then for other cases, go to another attribute and get the highest information. This algorithm is mainly user for finding the dependent attribute. Other methods not do this. References Class Notes, Book chapters, Class presentation materials.