Dynamic Integrated System for Detecting and Fixing Vulnerability Bugs

Received Jan 9, 2018 Revised Jan 27, 2018 Accepted Feb 7, 2018 Bugs are one of the important barriers in the field of software development . Concurrent and frequent bugs are non-deterministic in nature and in the time of vulnerability testing. It is difficult to detect the dynamic bugs with a high representation of vulnerability that causes the damage to the software products. Existing vulnerability testing methods relied on the conventional static testing techniques of finding and fixing the bugs but it does not show a high ratio of they handle or require specific hardware support. It does not include in the clustering approach. To overcome the limitations in the existing techniques, this proposed methods Modified Particle Swarm Optimization (MPSO), Expectation Maximization (EM) Clustering and Variable Neighborhood search. The primary input dataset is preprocessed to obtain the relevant and irrelevant data partition and optimized dataset was given as input to the Modified Particle Swarm Optimization (MPSO) technique Keyword:


INTRODUCTION
Software Quality Assurance is used to ensure the quality of software in the field of software product development.It will improve the standard of the out coming Software Bugs are one of the important factors in STLC (Software Testing Life Cycle).The increase of bugs will reduce the software quality.So bug detection or prediction will be helpful to the software developers and testers.It also includes the number of bugs, non-trivial bugs, number of major bugs, number of critical bugs, number of high priority bugs.Using the information the vulnerable part of the software can be identified.The identification will improve the software quality assurance.The fundamental idea is to gather insights portraying a program's runtime conduct over numerous executions.Clustering analysis concentrates on purely numerical data only.The typical clustering algorithms include the k-means, EM algorithm, and their variants.The bug rejection using to software testing maintained.The data mining is used to the clustering techno logies.They result in two algorithm execution on the software quality assurances testing and vulnerability detection on the work.

PROPOSED METHOD
The proposed methodology was the combination of three algorithms such as Modified Particle Swarm Optimization (MPSO), Expectation Maximization (EM) Clustering and Variable Neighborhood search.The input dataset is preprocessed to prepare the datasets to cluster the irrelevant data removed from the dataset, then the entropy values are calculated to get the maximum vulnerability level of test cases.The preprocessed dataset is split up into several subsets.Feature selection is used as next step to reduce the execution time of the data.Modified Particle Swarm Optimization (MPSO) is used to predict the bugs in eeffective way.Feature reduced data is split up into several clusters.For clustering Expectation Maximization (EM) clustering is used to cluster the bugs in repeated manner.Variable Neighborhood Search algorithm is used to find the sensitive data in the dataset.This sensitive data will make the software as more vulnerable.Vulnerability detection will definitely improve the software quality assurance.

ALGORITHM 1 MPSO Preliminaries and Assumptions FEATURE SELECTION (MODIFIED PARTICLE SWARM OPTIMIZATION) E entropy
Step1:fori=0: Ado ∑ Step 3:end for Step 4:forj=0: A.size-1do ∑ Step 6:end for Step 7:Ig information gain Step 8:Ba best attributes Step 9:forj=0: A.size-1do Step 10: if then Step 11:end if Step 12: end for The MPSO performs the head clustering and calculate the entropy values from the input datasets that is equally proportional to the sum of input vulnerability test case datasets.It include the size of values in entropy detection found in the first method, then the feature selection can be processed in the next step by using the EM technique.The number of holo-entropy and entropy of each vulnerability cluster heads was given as input into the EM to reduce the dynamic bugs by doing the iteration process with constrained threshold value, then it chooses the cluster data size and data clustering in maximum values.The maximum values of vulnerability test cases moves into the maximum queue.

√∑ ( ) if then end if end for end for
Search neighborhood vulnerability a value to maximum vulnerability test cases with threshold vulnerability value used by the system and returns the maximum and data minimum values to predict to the final daataset.

RESULT AND DISCUSSION
The software quality assurance checking on MPSO and EM algorithm is designed for reliable and effectivw bg predictions.It had been us ed to a number of clustering data.Figure 1 expresses the bug detection details with filteration reports.This result is testing with vulnerability detection from a cluster in software quality assurances.Bugfree software is having high quality.It also includes the number of bugs, non -trivial bugs, number of major bugs, number of critical bugs, number of high priority bugs in Figure 2. Using this information the vulnerable part of the software can be identified.This identification will improve the software quality assurance.The fundamental idea is to gather insights portraying a program's runtime conduct over numerous executions.

Figure 2. Bug detection reports
The method used Modified Particle Swarm Optimization (MPSO), Expectation Maximization (EM) Clustering and Variable Neighborhood search for vulnerability test cases.At preprocessing irrelevant data are removed from the dataset and moved into the EM mechanism.Then preprocessed dataset was split up into several subsets.Therein data cluster to the process will be done by the EM and vulnerability filter was increased drastically by this proposed work whiose details are explained in Figure 3.The quality assurances in volatility detection in the execution of this process had given the maximum output ranges when comparing to existing techniques.The process was improved to vulnerability detection because of three step process by this proposed method includes clustering of each test case heads and then entropy and holo -entropy values were moved to found the software quality check test in bug error correction .thechart explain to the improvement to the high-level process.First software data upload to the bug and clustering data then software quality assurances data detection.Finally, many solutions have been introduced to help fixing the dynamic vulnerability bugs.These all three dynamic methods that integrate dynamic bug detection and fixing comparing with static methods that generate that shows bugs again offline.This model have been simulated using Java 8 with dynamic test cases with different parameters shown the better bug detection and fixing and also controls both atomicity violations and order violations.

CONCLUSION
The proposed technique was shown the success rates in vulnerability bug rejection in software quality assurance environments by applying different of testing parameters.The process was tested with threshold values and vulnerability constraints at each stage.The combination of three algorithms is integrated such as Modified Particle Swarm Optimization (MPSO), Expectation Maximization (EM) Clustered and Variable Neighborhood search.It was preprocessing the irrelevant data and vulnerability test case validation.Then preprocessed dataset is split up into several subsets.Feature selection is used as next step to reduce the execution time of the data.The vulnerability test was detected with different parameters set and achieved the maximum bug reduction, that shows this proposed method for vulnerability testing experiments was executed in the software quality assurances test beds.In future this system will be tested in software product based vulnerability environments.

Figure 3 .
Figure 3. Vulnerability testing results using proposed method for Detecting and Fixing Vulnerability Bugs (R. Anitha) 109