Mohsen Beheshti

Diabetes Data Analysis and Prediction Model Discovery Using RapidMiner

Data mining techniques have been extensively applied in bioinformatics to analyze biomedical data. In this paper, we choose the Rapid-I¿s RapidMiner as our tool to analyze a Pima Indians Diabetes Data Set, which collects the information of patients with and without developing diabetes. The discussion follows the data mining process. The focus will be on the data preprocessing, including attribute identification and selection, outlier removal, data normalization and numerical discretization, visual data analysis, hidden relationships discovery, and a diabetes prediction model construction.

Discovering Decision Tree Based Diabetes Prediction Model

Data mining techniques have been extensively applied in bioinformatics to analyze biomedical data. In this paper, we choose the Rapid-I’s RapidMiner as our tool to discover decision tree based diabetes prediction model from a Pima Indians Diabetes Data Set, which collects the information of patients with and without developing diabetes. Following the data mining process, our discussion will focus on the data preprocessing, including attribute identification and selection, outlier removal, data normalization and numerical discretization, visual data analysis, hidden relationships discovery, and a diabetes prediction model construction.

The Computing Alliance of Hispanic-Serving Institutions: Supporting Hispanics at Critical Transition Points

Hispanics have the highest growth rates among all groups in the U.S., yet they remain considerably underrepresented in computing careers and in the numbers who obtain advanced degrees. Hispanics constituted about 7% of undergraduate computer science and computer engineering graduates and 1% of doctoral graduates in 2007--2008. The small number of Hispanic faculty, combined with the lack of Hispanic role models and mentors, perpetuates a troublesome cycle of underrepresentation in STEM fields. In 2004, seven Hispanic-Serving Institutions (HSIs) formed the Computing Alliance of Hispanic-Serving Institutions (CAHSI) to consolidate their strengths, resources, and concerns with the aim of increasing the number of Hispanics who pursue and complete baccalaureate and advanced degrees in computing areas. To address barriers that hinder students from advancing, CAHSI defined a number of initiatives, based on programs that produced promising results at one or more institutions. These included the following: a CS-0 course that focuses on adoption of a three-unit pre-CS course that uses graphics and animation to engage and prepare students who have no prior experience in computing; a peer mentoring strategy that provides an active, collaborative learning experience for students while creating leadership roles for undergraduates; an undergraduate and graduate student research model that emphasizes the deliberate and intentional development of technical, team, and professional skills and knowledge required for research and cooperative work; and a mentoring framework for engaging undergraduates in experiences and activities that prepare them for graduate studies and onto the professoriate. CAHSI plays a critical role in evaluating, documenting, and disseminating effective practices that achieve its mission. This paper provides an overview of CAHSI initiatives and describes how each addresses causes of underrepresentation of Hispanics in computing. In addition, it describes the evaluation and assessment of the initiatives and presents the results that support CAHSI’s claim of their effectiveness.

A Knowledge-Based System Implementation of Intrusion Detection Rules

This research determines the feasibility of using an Exsys Corvid based expert system to detect and respond to network threats and appropriately administrate a Linux-based iptables firewall in real-time. In our implementation, we attempt to replace the human domain expert required for creating the expert system knowledge base with intrusion detection rules created by data-mining on network traffic. Our expert system will be used in conjunction with intrusion detection classification rules provided by the See5 data-mining tool, which have, in turn, been created based on the data fusion of normal and malicious network traffic from multiple network sensors.

Improving Security through Analysis of Log Files Intersections

The paper discusses our research in development of general and systematic methods for intrusion prevention. The key idea is to use data mining techniques to discover repeated patterns of system features that describe program and user behavior. Server systems customarily write comprehensive activity logs whose value is useful in detecting intrusion. Unfortunately, production volumes overwhelm the capacity and manageability of traditional approach. This paper discusses the issues involving large-scale log processing that helps to analyze log records. Here, we propose to analyze intersections of firewall log files with application log files installed on one computer, as well as intersections resulting from firewall log files with application log files coming from different computers. Intersections of log files are substantially shorter than full logs and consist of records that indicate abnormalities in accessing single computer or set of computers. The paper concludes with some lessons we learned in building the system.

Data Fusion Support for Intrusion Detection and Prevention

Our main purpose for this work is to examine how to integrate multiple intrusion detection sensors in the order to minimize the number of incorrect-alarms The first problem is how to integrate data from multiple sensors, and the second how to identify most important data provided by multiple sensors. We are currently developing series of analytical models to use potential benefits of multiple sensors for reducing false alarms. The purpose of this presentation is to discuss implementation of prototype multisensor based intrusion detection system. We are especially interested in analyzing traffic that has an abnormal or malicious character and should prompt a closer look. A specific feature of the model is that the systems use multiple sensors to process log files

Work in progress - peer-led team learning implementation in computer science

The Computing Alliance of Hispanic Serving Institutions (CA-HSI) is a consortium of eight institutions that is committed to increasing the number of Hispanics who earn baccalaureate and advanced degrees in computing. CA-HSI is implementing and promoting the development of peer-led team learning (PLTL) in the computing curriculum as one of its interventions to increase the number of students who succeed in computing gatekeeper courses. PLTL is a proven strategy for retention and motivation of students having already shown its effectiveness in other disciplines. It utilizes student-driven focus groups to confront issues of lack of academic and social support. CA-HSI is creating a repository for support materials for PLTL implementation and it is also promoting implementation within the Alliance as well as other institutions. This paper discusses the materials and mechanisms for implementations, and it presents initial results of the intervention. At least 2,000 computing students have used these materials to date.

Analysis of Log Files Intersections for Security Enhancement

In this paper, we discuss our research in developing general and systematic methods for intrusion prevention. The key idea is to use data mining techniques to discover regular patterns of system features that describe program and user behavior. Server systems invariably write detailed activity logs whose value is useful in detecting intrusion. Unfortunately, production volumes overwhelm the capacity and manageability of traditional approach. This paper discusses the issues involving large-scale log processing that helps analyze log records. In this paper we propose to analyze intersections of log files that come from different applications and firewalls installed on one computer, and intersections resulting from log files coming from different computers. Intersections of log files are substantially smaller than full logs and consist of records that indicate abnormalities in accessing single computer or set of computers. The paper concludes with some lessons we learned in building the system

Component-Based Software Architecture Design for Network Intrusion Detection and Prevention System

A computer network intrusion detection and prevention system consists of collecting network traffic data, discovering user behavior patterns as intrusion detection rules, and applying these rules to prevent malicious and misuse. Many commercial off-the-shelf (COTS) products have been developed to perform each of these tasks. In this paper, the component-based software engineering approach is exploited to integrate these COTS products as components into a computerized system to automatically detect intrusion rules from network traffic data and prevent future potential attacks. The component-based software architecture of this kind of system is designed, COTS components are analyzed, adaptor components to compose COTS products are developed, and the system implementation is illustrated.

DAO-R: Integrating Data Aggregation and Offloading in Sensor Networks via Data Replication

We study overall storage overflow problem in sensor networks, wherein data-collecting base station is not available while more data is generated than available storage spaces in the entire network. Existing research designs a two-stage solution to solve this problem. It first aggregates overflow data to the size that can be accommodated by the available storage capacity in the network, and then offloads the aggregated data into the network to be stored. We refer to this naive two-stage solution as DAO-N. In this paper, we demonstrate that this approach does not necessarily achieve good performance. We propose a more unified method that is based upon data replication techniques, referred to as DAO-R, in order to improve the performance of DAO-N. Specifically, we design two energy-efficient data replication algorithms to integrate data aggregation and data offloading in DAO-N. We show via extensive simulations that DAO-R outperforms DAO-N by around 30% in terms of energy consumption under different network parameters.