Roshani Ade

Genetic algorithm based feature selection approach for effective intrusion detection system

Intrusion detection system (IDS) is the system which identifies malicious activity on the network. As the Internet volume is increasing rapidly, security against the real time attacks and their fast detection issues gain attention of many researchers. Data mining methods can be effectively applied to (IDS) to tackle the problems of dynamic huge network data and to improve IDS performance. We can reduce the time complexity by selecting only useful features to build model for classification. There are many features selection techniques are developed either to select the features or extract features. In this paper, an evolutionary approach for feature selection is proposed which is based on mathematical intersection principle. Genetic algorithm (GA) is used as a search method while selecting features from full NSL KDD data set along with the intersection principle of selecting those only who appears everywhere in the experiment. The results of proposed approach when compared using classifiers, it shows tremendous growth in accuracy of a Naïve Bayes classifier with reduced time and minimum number of features.

Classification of Students Using psychometric tests with the help of Incremental Naive Bayes Algorithm

this study, we validate that the incremental leaning as a technique in data mining can be used to classify the students according to their interest by conducting some aptitude test including psychometric tests on students. So that the students can get the correct carrier choice, student can learn the subject in which he/she is interested and improve their as well as institutes performance in terms of result. Recent years have observed very increasing interest in the topic of incremental learning, as it is having the ability to learn from new data introduces with the system even after the classifier has been produced from the formerly available data. It is required that the leaning should be done without accessing previously learned data and must remember previously acquired knowledge. This can be achieved by using incremental naive bayes classifier.

Software Requirement Engineering Risk Prediction Model

Analysis of many software program assignments from 2011 through 2014 shows interesting patterns. When you compare large assignments that have efficiently attained their own cost as well as schedule quotes versus the ones that ran delayed, ended up over spending budget, or maybe ended up, half a dozen popular difficulties ended up noticed: weak requirement analysis and management, weak cost calculating, weak handling of requirement change requests, weak milestone monitoring as well as requirement gold plating habit. By comparison, prosperous local software program assignments tended to be much better than on-site software development and management. Maybe the most interesting part of most of these many problem areas can be like everyone is coordinating project management instead of using technical focus. Author focused over the impact of software requirement change requests and requirement gold plating while dealing with on-site project assignments. Author also evaluated new algorithmic model to avoid global software engineering requirement failure, which in turn curtails the estimated time and budget with client satisfaction.

A Software Project Risk Analysis Tool Using Software Development Goal Modeling Approach

There are fewer practices for identification of risk factors in software development though everyone is aware of the impact of risks over project’s success parameters. To avoid risks or to identify risks, each team member needs to practice goal modeling at each phase of software development life cycle. It is not only project manager’s task to focus over risk occurrences but also it must be the duty of each team member to keep an eye on an assigned task and relative risk factors. As per an industrial need, “Sketch the Risk” tool is designed and developed in the most user-friendly way and hence, any team member from the project group can analyze given task from many views. The focus of development of “Sketch the Risk” tool is to identify risk events and avoid risk factors like project development delays, increase in the estimated cost, unnecessary rework, human resources utilization, etc. Unlike of existing approaches that are focusing over early stage to identify risk events, we proposed and developed new technique that takes care at each stage of software development life cycle. In essence, this paper presents identification goals and sub-goals in recurring manner just parallel to the spiral model of the software process.

Feature Selection based Classification using Naive Bayes, J48 and Support Vector Machine

One way to improve accuracy of a classifier is to use the minimum number of features. Many feature selection techniques are proposed to find out the most important features. In this paper, feature selection methods Co-relation based feature Selection, Wrapper method and Information Gain are used, before applying supervised learning based classification techniques. The results show that Support vector Machine with Information Gain and Wrapper method have the best results as compared to others tested. General Terms Feature selection; supervised learning

Logistic Regression Learning Model for Handling Concept Drift with Unbalanced Data in Credit Card Fraud Detection System

Credit card is the well-accepted manner of remission in financial field. With the rising number of users across the globe, risks on usage of credit card have also been increased, where there is danger of stealing credit card details and committing frauds. Traditionally, machine learning area has been developing algorithms that have certain assumptions on underlying distribution of data, such as data should have predetermined and fixed distribution. Real-word situations are different than this constrained model; rather applications often face problems such as unbalanced data distribution. Additionally, data picked from non-stationary environments are also frequent that results in the sudden drifts in the concepts. These issues have been separately addressed by the researchers. This paper aims to propose a universal framework using logistic regression model that intelligently tackles issues in the incremental learning for the assessment of credit risks.

Prediction of Student's Performance based on Incremental Learning

It is necessary to use Student dataset in order to analyze students performance for future improvements in study methods and overall curricular. Incremental learning methods are becoming popular nowadays since amount of data and information is rising day by day. There is need to update classifier in order to scale up learning to manage more training data. Incremental learning technique is a way in which data is processed in chunks and the results are merged so as to possess less memory. For this reason, in this paper, four classifiers that can run incrementally: the Naive Bayes, KStar, IBK and Nearest neighbor (KNN) have been compared. It is observed that nearest neighbor algorithm gives better accuracy compared to others if applied on Student Evaluation dataset which has been used.

Enhancing performance of keyword query over structured data

Keyword query applies to the database which accommodate structured data provide a search option over text attributes that uses a probability based ranking technique but query facing the issue of poor quality results. The keyword matches with multiple entities because the user does not provide exact data from which we can get better query results. Previous work analyze reasons behind the multiple answer and estimate the degree of difficulty for such tough query by using a ranking stability technique, but not addressed the problem of absent values under the attributes and it also imposes effectiveness problem of the query results. In this paper, we recognize and handle the issue of absent values by using both the inferring and retrieving based approaches. We evaluated our technique using real world data set and present performance advantages of our approach which improves the user satisfaction.

MLP-based undersampling technique for imbalanced learning

The imbalanced learning problem is becoming pervasive in todays data mining applications. This problem refers to the uneven distribution of instances among the classes which poses difficulty in the classification of rare instances. Several undersampling as well as oversampling methods were proposed to deal with such imbalance. Many undersampling techniques do not consider distribution of information among the classes, similarly some oversampling techniques lead to the overfitting or may cause overgeneralization problem. This paper proposes an MLP-based undersampling technique (MLPUS) which will preserve the distribution of information while doing undersampling. This reduction can be done on the basis of stochastic measure evaluation. Experiments are performed on 10 real world data sets for the evaluation of performance of MLPUS.

Patient Controlled Encryption using Key Aggregation

Cloud has become very important in internet world. Cloud provides storages, platforms which improves the functionality. Cloud storage shows how securely and flexibly we can store and share our data. This technique introduces a special type of encryption called as key-aggregate cryptosystem which allows user to share their data partially across cloud and which produces constant size ciphertext. In this technique user provide a constant-size aggregate key for different ciphertext classes in cloud storage, but the other encrypted files outside the class remain confidential. We also compare this technique with existing one. We implemented this cryptosystem for public-key patient-controlled encryption system.