Chapter 6 Outline

1.6 Outline of Chapters

Since the main concept of this whole research work revolves around identifying the image of ready to harvest crop it is necessary that the whole research area should be deep within the area of datamining and the principles related to it.

Since the problem is well understood like any other research work the primary requirement is to identify whether such a work as stated in problem definition has been already done or not. And if it is done, then to what extent the work has been done. Therefore, to understand the circumstances of the concern research work chapter 2 has been designed. Chapter 2 provides a detailed Literature Survey to provide as base to the research work. The chapter is broadly divided into three different

Or in simple terms it may be defined as the process for extracting or mining knowledge from large amounts of data.So the chapter 3 will provide a brief overview on the various datamining techniques. This is because there should be a clear understanding on the basic principles of the area of concern because any research is incomplete without the knowledge of the basic principles. So not only that the chapter gives the idea about data mining it also describes some of its techniques which leads to the ability to differentiate each one of them from other and allows one to explore the possibilities amongsts them. To do so the techniques studied in this chapter in brief are Association Rule, classification rule, Frequent Episodes and Deviation

But the area of data mining consists of a number of Image Classification Techniques. Therefore, in the next chapter i.e. chapter 5, analysis of various classification and clustering techniques is done to identify best suited technique for the problem of the research work stated.The techniques considered for comparison are Artificial Neural Network, Decision Tree, Support Vector Machine and K-Means Clustering. Artificial Neuron of ANN follow the action of the human brain neuron such that each neuron accepts the output of the neighbouring neuron, perform its own task and send the output to the next level of neurons. Decision tree is based on a hierarchical structure in which at each level a test is given to one or more attribute values so as to have one of two outputs. The classification between any two classes is done with the help of a hyperplane in case of a Support Vector Machine. K-Means Clustering is a simple algorithm with an easy procedure to follow. The word K here specifies the number of clusters created during the process of classification.Each algorithm has its own advantages and disadvantages. So this chapter in detail makes a comparative analysis amongsts these technique to identify the best suited for this resarch work. Chapter 5 has its own essence because not only that it is providing with the best suited technique for classification or