Vijay M. Wadhai

Evaluation of Decision Tree Pruning Algorithms for Complexity and Classification Accuracy

ABSTRACT Classification is an important problem in data mining. Given a database of records, each with a class label, a classifier generates a concise and meaningful description for each class that can be used to classify subsequent records. A number of popular classifiers construct decision trees to generate class models. These classifiers first build a decision tree and then prune subtrees from the decision tree in a subsequent pruning phase to improve accuracy and prevent “overfitting”. In this paper, the different pruning methodologies available & their various features are discussed. Also the effectiveness of pruning is evaluated in terms of complexity and classification accuracy by applying C4.5 decision tree classification algorithm on Credit Card Database with pruning and without pruning. Instead of classifying the transactions either fraud or non-fraud the transactions are classified in four risk levels which is an innovative concept.

An Adaptive parameter free data mining approach for healthcare application

In today’s world, healthcare is the most important factor affecting human life. Due to heavy work load it is not possible for personal healthcare. The proposed system acts as a preventive measure for determining whether a person is fit or unfit based on his/her historical and real time data by applying clustering algorithms viz. K-means and D-stream. Both clustering algorithms are applied on patient’s biomedical historical database. To check the correctness of both the algorithms, we apply them on patient’s current biomedical data. The Density-based clustering algorithm i.e. the D-stream algorithm overcomes drawbacks of K-means algorithm. By calculating their performance measures we finally find out effectiveness and efficiency of both the algorithms.

Application of Hippocratic principles for privacy preservation in social network

With the number of users of social network growing exponentially, the need of protecting the user privacy in network has gain the prime importance. While joining a social network, the user is requested to fill up a lot of unnecessary information like educational background, birth date, interests etc. This information may get leaked or mal-accessed if not protected with proper security measures. The data stored in social network may be attacked in different ways according to purpose of attack. In this paper we identify basic types of privacy breaches in social network. Secondly, we study the concept of Hippocratic principles. We propose a simple classification of the information requested from the user when he joins the social network. We also propose a privacy preserving model based on Hippocratic principles, specifically for Purpose, Limited Disclosure, Consent and compliance. Our proposed model work on privacy metadata, query analyzer is extended to check the define policy before giving the result out. This model can be used while mining on private data, which will help to enhance the privacy level of trustworthiness among internet users.

Constraint-based Web Log Mining for Analyzing Customers’ Behaviour

Analysis of Web logs is one of the important challenges to provide Web intelligent services. Association rule mining algorithms are used widely to track users web behaviour. Due to large amount of data many times the rules formed by these algorithms are very long and redundant. Recently Constraintbased mining approaches have received attention to deal with these big and redundant association rules. In this paper we discuss the Constraint based web mining approach used to reduce the size of association rules derived from Web log. The approach proves effective in reducing the overlap of information and also improves the efficiency of mining tasks. Constraint-based mining enables users to concentrate on mining their interested association rules instead of the complete set of association rules.

Concept Adapting Real-Time Data Stream Mining for Health Care Applications

Developments in sensors, miniaturization of low-power microelectronics, and wireless networks are becoming a significant opportunity for improving the quality of health care services. Vital signals like ECG, EEG, SpO2, BP etc. can be monitor through wireless sensor networks and analyzed with the help of data mining techniques. These real-time signals are continuous in nature and abruptly changing hence there is a need to apply an efficient and concept adapting real-time data stream mining techniques for taking intelligent health care decisions online. Because of the high speed and huge volume data set in data streams, the traditional classification technologies are no longer applicable. The most important criteria are to solve the real-time data streams mining problem with ‘concept drift’ efficiently. This paper presents the state-of-the art in this field with growing vitality and introduces the methods for detecting concept drift in data stream, then gives a significant summary of existing approaches to the problem of concept drift. The work is focused on applying these real time stream mining algorithms on vital signals of human body in health care environment.

Novel Threshold Formulation for Energy Detection Method to Efficient Spectrum Sensing in Cognitive Radio

Cognitive radio (CR) is an upcoming technology in wireless communications. Available spectrum used by licensed user is called as primary user (PU). Unused licensed spectrum is called spectrum hole. CR i.e., unlicensed user is called as secondary user (SU). The important task to the CR is to efficiently use the available spectrum without disturbing PU. The success of CR depends on spectrum sensing efficiency. In energy detection method, we have compared the energy of the samples with threshold (denoted as ( lambda )). If the sample energy is greater than the threshold, we say the signal is present; otherwise, the signal is absent. This method does not require any prior information of the signal, so energy detection is a popular sensing method due to its simplicity. Through this paper, we would like to propose a novel threshold formulation for energy detection method to efficient spectrum sensing in cognitive radio. Proposed formulation avoids Q inverse function; also, this formulation is independent of number of samples. With rigorous analysis of the proposed formulation for different SNRs, it gives better results compared to other systems.

Watermarking Relational Database Using Hindi Phonemes and Hill Cipher Technique

Digital watermarking now-a-days become more and more important due to tremendous availability of digital data on internet. The use of databases in various internet base applications has increased tremendously and theft of the data from database is a main concern for the database owners. Therefore, it is crucial to protect the piracy of the database. Most important technique in watermarking is its secret key and insertion location. In this paper, a new relational database watermarking method for non-numeric data is proposed based on nonnumeric attributes. A mark is computed based on the hill cipher technique. The position where mark is to be inserted is taken by the user. Our method is effective as it is robust and secure against different forms of malicious attacks.

Hippocratic social network

Social network is a network of people spread across all over the globe. Each social network user has a profile, which stores users personal information, his likes, interests etc. The number of social network users is growing exponentially, every day. This makes social network an ultimate repository of large user data and an important live information source. The large information available over social network attracts the attention of business, corporate and marketing people. So, these people try mining the user data/profile through different ways. Also, as most of the user profiles are publicly visible, it is very easy to obtain a particular users information without his concern. This leads to a privacy breach causing leakage of users private information, without even a hint of it to the user. We studied 100 facebook live users profiles and facebook privacy policy, to understand the privacy awareness in facebook users. In this paper, we present results of the surveys conducted in this study. We, further, propose a new generic framework named `Hippocratic Social Network, to enhance the personal level privacy in facebook and other online social networking sites.

Dynamic data mining approach to WMRHM

Developments in sensors, miniaturization of low-power microelectronics, and wireless networks are becoming a significant opportunity for improving the quality of health care services. Since the population is growing, the need for high quality and efficient healthcare, both at home and in hospital, is becoming more important. This paper presents the innovative wireless sensor network based Mobile Real-time Health care Monitoring (WMRHM) framework which has the capacity of giving health predictions online based on continuously monitored real time vital body signals. Our approach focused towards handling all kinds of vital signals like ECG, EMG, SpO2 etc. which previous work was not supporting. While predictions the framework considers all parameters like patient history, domain experts rules and continuously monitored real-time signals. Implementation and results of applying clustering algorithms (Graph theoretic, K-means) on patients historical health data for forming the health rule base are discussed here. The framework has been designed to perform the analysis on the instantaneous and stream (continuous) data over a sliding time window which applies dynamic data mining on the live data. The comparative analysis on vital signals made from various clustering algorithms adds extra dimension to global risk alerts and help doctors to diagnose more accurately.

A Comparative Study of Traditional and Mobile based ECG System Algorithms

ECG is a very important tool for cardiac health status measurement and detection of various diseases at their early stage. Thus, it will be helpful in providing such analysis on the move using a mobile device. The basic aim for such implementation would be an accurate, efficient yet lightweighted algorithm that is simple to implement on mobile. With a lot of traditional techniques been developed, it is important to study them correctly and use the right combination to achieve the desired result for efficient processing. This paper discusses comparative study of traditional and mobile based ECG feature extraction techniques and design of a mobile based ECG classification system. General Terms Your general terms must be any term which can be used for general classification of the submitted material such as Pattern Recognition, Security, Algorithms et. al.