Week 5: Data mining

Data mining is powerful tool in finding large set of data. The article discuss about techniques of data mining such as Decision tree, Neural network and CART methodology. Comparing all the three techniques it seems that adoption of decision tree is the most simple technique, but it is not as useful as neural network.In case of decision trees result can be derived according to the design of the tree, so if there is no adequate design of the tree than the result will be different. The back point of decision tree is, one decision tree can find out one patter. There are different splitting rules which could lead to different number of terminal tree nodes, that may also lead to different results from the same data.

A part from decision tree, neural network is one of the other data mining techniques which has been frequently used. Neural network could assist experts in credit card risk assessment, it is generated as a black box; users just feed the training data to construct it and testing data to exam it. The major difference between both the techniques is decision tree are simple and easy to understand, but on other side neural network is bit complicated. With Neural Network, a considerable amount of time and money need to be spent in training as it would require the analyst to have background understanding of Neural Network such as:

How to best design its algorithm ?
How to maintain to get the accurate outcome?

From the article it is seen that both the tools had been widely adopted by the organization but than also it is necessary to have a further research on both the techniques to get an accurate results.