Machine learning with classification can effectively be applied for many applications, especially those with complex measurements. Therefore classification technique can be used for prediction of diseases like cancer, liver disorders and heart disease etc which involve complex measurements. This is part of growing demand and much interesting towards predictive diagnosis. It has also been established that classification and learning methods can be used effectively to improve the accuracy of prediction of a diseases and its recurrence. General TermsMachine learning, support vector machine, random forest, kernels, linear, non linear, radial basis function and sigmoid.

This paper evaluates a system or system component, by manual or automated means to verify that the system satisfies specified set of requirements. The Test Management and Control involves three major activities; test planning, test execution and defect management which needs to be planned and monitored in a structured manner to ensure delivery of quality product adhering to all project timelines. The purpose of this document is to present the methods, procedures, and approach to be used in the verification and validation for Test Management & Control.It consists of important activities like defining all the testing methodologies in detail for the test planning and lay down a process and procedure so that test plan will reflect all the activities which are going to get executed during the execution phase. The Test execution phase is the actual verification phase where you need to monitor the test case execution closely, and make sure that you are well within the defined Upper Control Limit (UCL) and Lower Control Limit (LCL). During the test execution phase which is clearly planned during the test planning phase, any deviation from these UCL & LCL requires a revised plan so that the deliverables are not going to jeopardy. So once we have clearly defined all the rules of the game during the planning phase and work on the same during the execution phase, we can certainly make sure that the product delivered will of high quality product. Managing Defects is also one of the very important phases all along the testing cycle. We need to pay extra attention to the defects getting fixed during the last few days of the testing cycle, so as to maintain the stability of the product. Do not fix the minor defects at the end of the testing cycle, unless they are critical and severe.

No abstract

The advancement of 4th Generation Languages (4GLs) brought the mutiny in the Information technology by making users free from the burden of developing logic for solution of the application problem. At the same time, it enhanced the need for implementation of ameliorated strategies to evaluate the intermediate operations. The development of 4GLs freed the users from the strained design of 'how to do it' for solutions of their applications. At the back end, it increased the complexity of transformation from non-procedural query to procedural query. The query processing and optimization manages these transformations, by transforming 4GLs query into appropriate relational algebra/calculus expression in the first phase, minimizing the operations in the next phase and then evaluating the expression through the efficient strategies for such operations in the third phase. The transformation to relational algebra/calculus expression and minimizing operations are straightforward processes. The development of ameliorated strategies depends on the file organizations used in the storing relations of the database (Secondary storage structures). This paper discusses the strategies designed for select operation, when tuples of operand relations are stored in one of various file organizations in the database. Here, an attempt is made to design ameliorated algorithms to evaluate select operation when the tuples of the operand relations are stored in any of the file organizations like (i) sequential file (ii) indexed sequential files with B-tree file having number of key field indexes in each node and B+ tree file organizations for multiple indexes where data pointers are stored in each leaf node of the multilevel indexes of the file organization. The literature available suggests that, there are good methods available when key fields are primary keys. The authors have not noticed any clear cut strategies, when key fields for indexing are non primary keys (non candidate keys). Here, an attempt is made to develop the strategies, when key fields are non-candidate keys.

No abstract