Current Issue ( Vol. : 8, Issue: 2, April-June 2020)   


Title: A LITERATURE REVIEW ON INTRUSION DETECTION SYSTEM USING KDDCUP’99 DATASET
Author: Kusum Lata, Mr. Manoj Yadav and Mr. Kailash Patidar
Keyword: IDS, Security Threats, Integrity, KDDCUP99
Page No: 23-33
Abstract: An Internet is extensively used technology for the data communication now-a-days but during the communication of data network can be influenced by rigorous type of security threats and security issues which may lead the loss of data or can corrupt the data. So, to preserve the integrity or confidentiality of personal data or information a system is developed is known an intrusion detection system which collect and examine the various areas from which network can get trap. For the detection of intrusion different methodologies has been developed. In this paper, literature of some of the techniques to identify and thwart from the intrusion and their detection techniques is discussing using KDDCUP99 dataset and also presents its merits and demerits.Download PDF


Title: AN ANALYSIS ON SOFTWARE QUALITY PROPHECY USING DATA MINING AND MACHINE LEARNING PROFICIENCY
Author: Shubha Dubey
Keyword: Software quality prophecy, metrics, machine learning, prophecy proficiency.
Page No: 34-40
DOI: 10.30876/JOHR.8.2.2020.34-40
Abstract: The need of the time is highly reliable software usage in the systems. However, time-consuming costly development processes to assure reliability. One lucrative approach is to aim reliability-enhancem ent activities to those modules that are likely to have the most problems. If in each module software quality prophecy models can predict the number of faults at the earliest it would be expected enough effective for reliability enhancement. The main purpose of this paper is to help developers categorize defects based on existing software metrics using data mining proficiencies and by this means improve the software quality. In this paper, various software quality prophecy strategies based on Bayesian belief network, neural networks, fuzzy logic, support vector machine and case-based reasoning are juxtapose. This study gives better relative understanding about these strategies, and helps to choose an approach based on existing resources software quality prophecy using software metrics in the literature.Download PDF


Title: AN ANALYSIS ON MACHINE LEARNING BASED PREDICTION FOR SUCCESS RATE OF PROJECT
Author: Vikas
Keyword: Forest, Logistic Regression, Prediction, Software Project, MAE, MSE, Precision,
Page No: 41-49
Abstract: The prediction of project success the exhaustive goal of various Industries. Whereas, it becomes more critical to execute the project successfully. To predict the project success various data mining a nd machine learning techniques such as k-nearest neighbor, SVM classier, logistic regression has been developed but, in this work, we use random forest classifier for the prediction of project success with reduced cost and schedule. The random forest classifier selects the dataset randomly from the available dataset and the generate the decision tree of the selected dataset and then apply the voting on the prediction results and whose score and accuracy will be maximum that will indicates the success of project. For the sample dataset we use online resource of kaggle and the experimental results is generated from the widely used machine learning programming language Python which helps in the analysis of the proposed methodology. The performance of proposed methodology is measured using the parameters such as Score, accuracy, precision, recall value, F1 score, mean absolute error and mean square error. The comparative analysis of the proposed methodology is done among the existing approach K-neighbor and Logistic regression. The score and accuracy value of our proposed methodology is 78% while other is less. Similarly, the F1 score, precision and recall value of proposed methodology is 43%, 50% and 42% while the K –neighbor and logistic regression is comparatively very less. Similarly, the comparative analysis of proposed and existing approach is done using mean absolute error and mean square error and the value is 22% and 11% which is very less. These results of proposed methodology improve the success rate of software project Download PDF