Forecasting weather and forewarning the public with weather alerts help save lives and property i... more Forecasting weather and forewarning the public with weather alerts help save lives and property immensely. Audio transmissions for alerting are not prone to local infrastructure failure and hence the most dependable. Reception of such alerts should have good clarity of audio. The Mean Opinion Score (MOS) method has traditionally been used in telephony to monitor audio reception quality. It requires an objective skilled listener using a scoring system for feedback used in improving audio quality. However, in weather radio, there is no dedicated listener 24/7. Without listening skills and listener training the MOS scoring could be skewed and of little use. Our paper focuses on the design, building and testing of audio-quality forensics system (AQFS) that uses measurements of network, radio and equipment parameters to monitor audio quality instead of listener feedback. Transmitter stations today have multiple connectivity options and the AQFS can alert the transmission operators to switch to alternate technology during failures. The AQFS also has control features to perform this switching automatically that we have tested. In this paper, we also report our findings, test results and inferences with recommendations. Apart from providing basic but detailed audio quality analysis and comparison of base and transmitted audio using two different mechanisms, the AQFS helped us detect complete audio loss, intermittent audio outages, Specific Area Message Encoding (SAME) Headers, audio anomalies, fades, drops and strange tones which correlate with fading and outage events. The analysis helped at problem transmitter sites in improving audio quality. INDEX TERMS 4G communication, audio over IP, broadcasting, mean opinion score, real time protocol, weather alert.
This paper looks at the impact of changing Spark's configuration parameters on machine learning a... more This paper looks at the impact of changing Spark's configuration parameters on machine learning algorithms using a large dataset-the UNSW-NB15 dataset. The environmental conditions that will optimize the classification process are studied. To build smart intrusion detection systems, a deep understanding of the environmental parameters is necessary. Specifically, the focus is on the following environmental parameters: the executor memory, number of executors, number of cores per executor, execution time, as well as the impact on statistical measures. Hence, the objective was to optimize resource usage and minimize processing time for Decision Tree classification, using Spark. This shows whether additional resources will increase performance, lower processing time, and optimize computing resources. The UNSW-NB15 dataset, being a large dataset, provides enough data and complexity to see the changes in computing resource configurations in Spark. Principal Component Analysis was used for preprocessing the dataset. Results indicated that a lack of executors and cores result in wasted resources and long processing time. Excessive resource allocation did not improve processing time. Environmental tuning has a noticeable impact.
Association rule mining, originally developed by , is a well-known data mining technique used to ... more Association rule mining, originally developed by , is a well-known data mining technique used to find associations between items or itemsets. In today's big data environment, association rule mining has to be extended to big data. The Apriori algorithm is one of the most commonly used algorithms for association rule mining . Using the Apriori algorithm, we find frequent patterns, that is, patterns that occur frequently in data. The Apriori algorithm employs an iterative approach where k-itemsets are used to explore (k + 1) itemsets. To find the frequent itemsets, first the set of frequent 1-itemsets are found by scanning the database and accumulating their counts. Itemsets that satisfy the minimum support threshold are kept. These are then used to find the frequent 2-itemsets. This process goes on until the newly generated itemset is an empty set, that is, until there are no more itemsets that meet the minimum support threshold. Then the itemsets are checked against a minimum confidence level to determine the association rules. The process of generating the frequent itemsets calls for repeated full scans of the database, and in this era of big data, this is a major challenge of this algorithm. Figure presents a flow chart of how the Apriori algorithm works. Traditional association rule mining algorithms, like Apriori, mostly mine positive association rules. Positive association rule mining finds items that are positively related to one another, that is, if one item goes up, the related item also goes up. Though the classic application of positive association rule mining is market basket analysis, applications of
International Journal of Advanced Computer Science and Applications, 2021
This paper commences by introducing the essentials of blockchain technology and then goes into ho... more This paper commences by introducing the essentials of blockchain technology and then goes into how Ethereum blockchain revolutionized blockchain. Smart contracts are presented in the context of showing how they play an important role in implementing rules regarding the Ethereum blockchain, allowing the user to regulate digital assets. The standards used in the Ethereum blockchain to build Non-Fungible Tokens (NFTs) are discussed. The paper concludes by presenting the benefits of NFTs as well as the use of Ethereum blockchain for future applications.
Object-oriented database systems began developing in the mid-80's out of a necessity to meet the ... more Object-oriented database systems began developing in the mid-80's out of a necessity to meet the requirements of applications beyond the data processing applications which were [are] served by relational database systems. This paper serves as an overview on the achievements of object-oriented database technology so far, and also discusses the weaknesses that have to be yet resolved by the object-oriented database community before object-oriented database technology can become as widespread as relational databases.
Transactions on Networks and Communications, Apr 23, 2021
In a shared Elasticsearch environment it can be useful to know how long a particular query will t... more In a shared Elasticsearch environment it can be useful to know how long a particular query will take to execute. This information can be used to enforce rate limiting or distribute requests equitably among multiple clusters. Elasticsearch uses multiple Lucene instances on multiple hosts as an underlying search engine implementation, but this abstraction makes it difficult to predict execution with previously known predictors such as the number of postings. This research investigates the ability of different pre-retrieval statistics, available through Elasticsearch, to accurately predict the execution time of queries on a typical Elasticsearch cluster. The number of terms in a query and the Total Term Frequency (TTF) from Elasticsearch’s API are found to significantly predict execution time. Regression models are then built and compared to find the most accurate method for predicting query time.
In this paper, a new technique for camera calibration using only GPS data is presented. A new way... more In this paper, a new technique for camera calibration using only GPS data is presented. A new way of tracking objects that move on a plane in a video is achieved by using the location and size of the bounding box to estimate the distance, achieving an average prediction error of 5.55m per 100m distance from the camera. This solution can be run in real-time at the edge, achieving efficient inference in a low-powered IoT environment while also being able to track multiple different vessels. Ship avoidance systems are typically based on radio signals. In this paper, a new technique for camera calibration using only GPS data is presented. To get a better position estimate from an object detector, the property that ships move on a plane is utilized. Also, a new way of tracking objects that move on a plane in a video is achieved by using the location and size of the bounding box to estimate the distance. This approach has many possible use cases in edge computing. This includes tracking airplanes in airports, ships in harbors, cars in parking lots, or humans in shopping malls. It also demonstrates how to efficiently track one specific object from other similar instances of that object. This can play an important role in traffic systems. For example, it is possible to keep track of free spaces in a parking lot. This information could be utilized in various IoT applications helping citizens to find a free parking space. Another use case is smarter traffic lights, which could use a camera to react to oncoming traffic quicker. For this, all that is needed for camera calibration and training or evaluating the object detector is 33 minutes of videos and ground truth GPS positions or something comparable. The methods utilized in this paper are fast enough to be deployed on inexpensive edge computing devices that can be mass-produced. a
Transactions on Machine Learning and Artificial Intelligence, Aug 1, 2020
This survey performs a thorough enumeration and analysis of existing methods for data stream proc... more This survey performs a thorough enumeration and analysis of existing methods for data stream processing. It is a survey of the challenges facing streaming data. The challenges addressed are preprocessing of streaming data, detection and dealing with concept drifts in streaming data, data reduction in the face of data streams, approximate queries and blocking operations in streaming data.
Supervisory Control and Data Acquisition (SCADA) systems and Industrial Control Systems (ICS) are... more Supervisory Control and Data Acquisition (SCADA) systems and Industrial Control Systems (ICS) are critical for infrastructure operations, production processes, automation systems, and other automated control systems. SCADA systems are vulnerable to cyber physical attacks from many vectors. The attack surface varies widely. The problem is complicated by the inherent focus on performance, reliability, and safety rather than cybersecurity. The Secure SCADA Framework establishes a security engineered approach evolving SCADA and ICS systems towards a more cybersecure posture. The encapsulating and integrating concepts of the Secure SCADA Framework require development and analysis of implementations achieving the eight framework goals. This work provides a solution for research, design, development, and evaluation of components of a Secure SCADA System. A Security Engineered SCADA Laboratory requires a test bed with which engineering and science designs can be developed. This paper presents a test bed which can model and collect data on simulated attacks, analyze data to evaluate performance, and provide the foundations for a Security Engineered SCADA Laboratory.
Journal of Modern Applied Statistical Methods, May 1, 2005
Provided in this article is an algorithm and code for computing exact critical values (or percent... more Provided in this article is an algorithm and code for computing exact critical values (or percentiles) for Friedman's nonparametric rank test for k related treatment populations using Visual Basic (VB.NET). This program has the ability to calculate critical values for any number of treatment populations ( k ) and block sizes ( ) b at any significance level ( ) α . We developed an exact critical value table for 2(1)5 k = and 2(1)15 b = . This table will be useful to practitioners since it is not available in standard nonparametric statistics texts. The program can also be used to compute any other critical values.
The Random Forest (RF) algorithm, originally proposed by Breiman et al. (1), is a widely used mac... more The Random Forest (RF) algorithm, originally proposed by Breiman et al. (1), is a widely used machine learning algorithm that gains its merit from its fast learning speed as well as high classification accuracy. However, despite its widespread use, the different mechanisms at work in Breiman's RF are not yet fully understood, and there is still ongoing research on several aspects of optimizing the RF algorithm, especially in the big data environment. To optimize the RF algorithm, this work builds new ensembles that optimize the random portions of the RF algorithm using genetic algorithms, yielding Random Genetic Forests (RGF), Negatively Correlated RGF (NC-RGF), and Preemptive RGF (PFS-RGF). These ensembles are compared with Breiman's classic RF algorithm in Hadoop's big data framework using Spark on a large, high-dimensional network intrusion dataset, UNSW-NB15.
International Journal of Changes in Education, Jan 8, 2024
The goal of this research study was to investigate the effects of D.A. Kolb's active experimentat... more The goal of this research study was to investigate the effects of D.A. Kolb's active experimentation stage of experiential learning theory (ELT) on teacher professional development. This qualitative interpretive phenomenological study addressed the problem of teachers' desires for effective professional learning experiences to improve student learning and instruction in the classroom. The research questions explored the perceptions of private school teachers regarding professional development sessions, specifically in the context of active experimentation. The study aimed to investigate the advantages of promoting a professional culture of continuous learning and teamwork among teachers through various methods such as lab classrooms, peer observation, professional learning communities, and peer coaching. The study adopts a descriptive approach utilizing focus group discussion and semi-structured interviews for data collection. Purposeful sampling was used to determine the sample and criterion sampling was used to select the participants. Biographic data were collected to provide background information about the participants involved. The biographic data allowed for stratification enabling the examination of various factors connected to the aim of the study. The study focused on understanding the benefits of providing teacher support through active experimentation during professional development sessions. Results showed that although teachers did not explicitly reference ELT when discussing effective strategies during professional development sessions, they referenced the value of hands-on, active participation in realclassroom stimulations during teacher professional development. The study's findings also indicate a need for creating a consistent framework for professional development sessions that includes active learning, teacher collaboration, and continuous support in a culture anchored in trust.
Advances in systems analysis, software engineering, and high performance computing book series, 2015
In this chapter, we use MySQL Database Cluster to demonstrate and discover the capabilities of ke... more In this chapter, we use MySQL Database Cluster to demonstrate and discover the capabilities of key based database sharding and provide the implementation details to build a key based sharded database system. After the implementation section, we present some examples of datasets that were sharded using our implementation. The sharded data is then used for data mining, specifically association rule mining. We present the results (association rules) for the sharded data as well as the non-sharded data.
Object-oriented database systems began developing in the mid-80’s out of a necessity to meet the ... more Object-oriented database systems began developing in the mid-80’s out of a necessity to meet the requirements of applications beyond the data processing applications which were [are] served by relational database systems. This paper serves as an overview on the achievements of object-oriented database technology so far, and also discusses the weaknesses that have to be yet resolved by the object-oriented database community before object-oriented database technology can become as widespread as relational databases.
Journal of Information Technology Research, Mar 25, 2022
With the steady rise in the use of smartphones, specifically Android smartphones, there is an ong... more With the steady rise in the use of smartphones, specifically Android smartphones, there is an ongoing need to build strong intrusion detection systems to protect ourselves from malicious software attacks. This work focuses on a subgroup of android malware, scareware. The novelty of this work lies in being able to detect the various scareware families individually using a small number of network attributes, determined by a recursive feature elimination process based on information gain. No work has yet been done on analyzing the scareware families individually. Results of this work show that the number of bytes initially sent back and forth, packet size, amount of time between flows and flow duration are the most important attributes that would be needed to classify a scareware attack. Three classifiers, Decision Tree, Naïve Bayes, and OneR, were used for classification. The highest average classification accuracy (79.5%) was achieved by the Decision Tree classifier with a minimum of 44 attributes.
International Journal of Information and Communication Technology Education, 2016
The goal of this research was to investigate the effects of automated testing software on levels ... more The goal of this research was to investigate the effects of automated testing software on levels of student reflection and student performance. This was a self-selecting, between subjects design that examined the performance of students in introductory computer programming classes. Participants were given the option of using the Web-CAT software-testing tool to evaluate their computer code. Student self-reported levels of reflection were measured using Kember et al.'s (2000) reflective thinking survey.
International Journal on Engineering, Science and Technology
Computer Science drives innovation and discovery, and Computer Science classes in High Schools, p... more Computer Science drives innovation and discovery, and Computer Science classes in High Schools, particularly AP Computer Science Principles, have also been linked to better performance in other High School courses, but a Code.org report shows that the state of Florida, particularly the Northwest Florida region, is far behind the rest of the nation in offering Computer Science classes at the High School level. This paper discusses strategies that are can be implemented to improve High School Computer Science offerings in like-regions that have lower Computer Science offerings. The paper also discusses the challenges that come up in addressing this issue and presents potential avenues for addressing these challenges. One major challenge that is addressed is creating the culture of inclusive computing by getting more women and under-represented groups interested in taking Computer Science courses. Though this is a Northwest Florida case study, most of the strategies discussed in this p...
High-altitude ballooning, along with other aerospace endeavors, requires extensive preplanning an... more High-altitude ballooning, along with other aerospace endeavors, requires extensive preplanning and preparation forvehicle launching. In ballooning specifically, weather conditions are especially effective and driving whether or nota launch can occur. As most flights must be shaped around the flight path, both for safety and recovery reasons,it is imperative that any acceptable flight path and day may be considered. The goal of this project is to minimize,using machine learning, the complexity and manpower requirements for determining if a launch can occur.
The Random Forest (RF) algorithm, originally proposed by Breiman [7], is a widely used machine le... more The Random Forest (RF) algorithm, originally proposed by Breiman [7], is a widely used machine learning algorithm that gains its merit from its fast learning speed as well as high classification accuracy. However, despite its widespread use, the different mechanisms at work in Breiman's RF are not yet fully understood, and there is still ongoing research on several aspects of optimizing the RF algorithm, especially in the big data environment. To optimize the RF algorithm, this work builds new ensembles that optimize the random portions of the RF algorithm using genetic algorithms, yielding Random Genetic Forests (RGF), Negatively Correlated RGF (NC-RGF), and Preemptive RGF (PFS-RGF). These ensembles are compared with Breiman's classic RF algorithm in Hadoop's big data framework using Spark on a large, high-dimensional network intrusion dataset, UNSW-NB15.
The K-Nearest Neighbor! (KNN) algorithm is one of the most widely used algorithms in data mining ... more The K-Nearest Neighbor! (KNN) algorithm is one of the most widely used algorithms in data mining for classification and prediction. The algorithm has several applications: in facial detection when used with deep learning, in biometric security applications etc. The traditional KNN algorithm involves an iterative process of computing the distance between a test data point and every data point in the training dataset, and classifying the object based on the closest training sample. This method first selects K nearest training data points for classifying a test data point and then predicts the test sample's class based on the majority class among those neighbors. If both the train and test datasets are large, this conventional form can be considered computationally expensive. Reduction of the massive calculation that is required to predict a data vector was our main goal, and with this intention, the training dataset was split into several buckets. The KNN algorithm was then performed inside a bucket, instead of iterating over the whole training dataset. We used the Jaccard Coefficient to determine the degree of similarity of a data vector with some arbitrarily defined data points P and placed similar data points in the same bucket. This was the core functionality of our hash function. The hash function determines the bucket number where the similar data vectors will be placed. Unlike the standard hashing algorithm, our approach of hashing was to maximize the probability of the hash collision to preserve the locality sensitiveness. Both the conventional and proposed methods were implemented in Hadoop's MapReduce framework. Hadoop gives us an architecture for handling large datasets on a computer cluster in a distributed manner and gives us massive scalability. The use of the locality sensitive hashing in KNN in Hadoop's MapReduce environment took less time than conventional KNN to classify a new data object.
Forecasting weather and forewarning the public with weather alerts help save lives and property i... more Forecasting weather and forewarning the public with weather alerts help save lives and property immensely. Audio transmissions for alerting are not prone to local infrastructure failure and hence the most dependable. Reception of such alerts should have good clarity of audio. The Mean Opinion Score (MOS) method has traditionally been used in telephony to monitor audio reception quality. It requires an objective skilled listener using a scoring system for feedback used in improving audio quality. However, in weather radio, there is no dedicated listener 24/7. Without listening skills and listener training the MOS scoring could be skewed and of little use. Our paper focuses on the design, building and testing of audio-quality forensics system (AQFS) that uses measurements of network, radio and equipment parameters to monitor audio quality instead of listener feedback. Transmitter stations today have multiple connectivity options and the AQFS can alert the transmission operators to switch to alternate technology during failures. The AQFS also has control features to perform this switching automatically that we have tested. In this paper, we also report our findings, test results and inferences with recommendations. Apart from providing basic but detailed audio quality analysis and comparison of base and transmitted audio using two different mechanisms, the AQFS helped us detect complete audio loss, intermittent audio outages, Specific Area Message Encoding (SAME) Headers, audio anomalies, fades, drops and strange tones which correlate with fading and outage events. The analysis helped at problem transmitter sites in improving audio quality. INDEX TERMS 4G communication, audio over IP, broadcasting, mean opinion score, real time protocol, weather alert.
This paper looks at the impact of changing Spark's configuration parameters on machine learning a... more This paper looks at the impact of changing Spark's configuration parameters on machine learning algorithms using a large dataset-the UNSW-NB15 dataset. The environmental conditions that will optimize the classification process are studied. To build smart intrusion detection systems, a deep understanding of the environmental parameters is necessary. Specifically, the focus is on the following environmental parameters: the executor memory, number of executors, number of cores per executor, execution time, as well as the impact on statistical measures. Hence, the objective was to optimize resource usage and minimize processing time for Decision Tree classification, using Spark. This shows whether additional resources will increase performance, lower processing time, and optimize computing resources. The UNSW-NB15 dataset, being a large dataset, provides enough data and complexity to see the changes in computing resource configurations in Spark. Principal Component Analysis was used for preprocessing the dataset. Results indicated that a lack of executors and cores result in wasted resources and long processing time. Excessive resource allocation did not improve processing time. Environmental tuning has a noticeable impact.
Association rule mining, originally developed by , is a well-known data mining technique used to ... more Association rule mining, originally developed by , is a well-known data mining technique used to find associations between items or itemsets. In today's big data environment, association rule mining has to be extended to big data. The Apriori algorithm is one of the most commonly used algorithms for association rule mining . Using the Apriori algorithm, we find frequent patterns, that is, patterns that occur frequently in data. The Apriori algorithm employs an iterative approach where k-itemsets are used to explore (k + 1) itemsets. To find the frequent itemsets, first the set of frequent 1-itemsets are found by scanning the database and accumulating their counts. Itemsets that satisfy the minimum support threshold are kept. These are then used to find the frequent 2-itemsets. This process goes on until the newly generated itemset is an empty set, that is, until there are no more itemsets that meet the minimum support threshold. Then the itemsets are checked against a minimum confidence level to determine the association rules. The process of generating the frequent itemsets calls for repeated full scans of the database, and in this era of big data, this is a major challenge of this algorithm. Figure presents a flow chart of how the Apriori algorithm works. Traditional association rule mining algorithms, like Apriori, mostly mine positive association rules. Positive association rule mining finds items that are positively related to one another, that is, if one item goes up, the related item also goes up. Though the classic application of positive association rule mining is market basket analysis, applications of
International Journal of Advanced Computer Science and Applications, 2021
This paper commences by introducing the essentials of blockchain technology and then goes into ho... more This paper commences by introducing the essentials of blockchain technology and then goes into how Ethereum blockchain revolutionized blockchain. Smart contracts are presented in the context of showing how they play an important role in implementing rules regarding the Ethereum blockchain, allowing the user to regulate digital assets. The standards used in the Ethereum blockchain to build Non-Fungible Tokens (NFTs) are discussed. The paper concludes by presenting the benefits of NFTs as well as the use of Ethereum blockchain for future applications.
Object-oriented database systems began developing in the mid-80's out of a necessity to meet the ... more Object-oriented database systems began developing in the mid-80's out of a necessity to meet the requirements of applications beyond the data processing applications which were [are] served by relational database systems. This paper serves as an overview on the achievements of object-oriented database technology so far, and also discusses the weaknesses that have to be yet resolved by the object-oriented database community before object-oriented database technology can become as widespread as relational databases.
Transactions on Networks and Communications, Apr 23, 2021
In a shared Elasticsearch environment it can be useful to know how long a particular query will t... more In a shared Elasticsearch environment it can be useful to know how long a particular query will take to execute. This information can be used to enforce rate limiting or distribute requests equitably among multiple clusters. Elasticsearch uses multiple Lucene instances on multiple hosts as an underlying search engine implementation, but this abstraction makes it difficult to predict execution with previously known predictors such as the number of postings. This research investigates the ability of different pre-retrieval statistics, available through Elasticsearch, to accurately predict the execution time of queries on a typical Elasticsearch cluster. The number of terms in a query and the Total Term Frequency (TTF) from Elasticsearch’s API are found to significantly predict execution time. Regression models are then built and compared to find the most accurate method for predicting query time.
In this paper, a new technique for camera calibration using only GPS data is presented. A new way... more In this paper, a new technique for camera calibration using only GPS data is presented. A new way of tracking objects that move on a plane in a video is achieved by using the location and size of the bounding box to estimate the distance, achieving an average prediction error of 5.55m per 100m distance from the camera. This solution can be run in real-time at the edge, achieving efficient inference in a low-powered IoT environment while also being able to track multiple different vessels. Ship avoidance systems are typically based on radio signals. In this paper, a new technique for camera calibration using only GPS data is presented. To get a better position estimate from an object detector, the property that ships move on a plane is utilized. Also, a new way of tracking objects that move on a plane in a video is achieved by using the location and size of the bounding box to estimate the distance. This approach has many possible use cases in edge computing. This includes tracking airplanes in airports, ships in harbors, cars in parking lots, or humans in shopping malls. It also demonstrates how to efficiently track one specific object from other similar instances of that object. This can play an important role in traffic systems. For example, it is possible to keep track of free spaces in a parking lot. This information could be utilized in various IoT applications helping citizens to find a free parking space. Another use case is smarter traffic lights, which could use a camera to react to oncoming traffic quicker. For this, all that is needed for camera calibration and training or evaluating the object detector is 33 minutes of videos and ground truth GPS positions or something comparable. The methods utilized in this paper are fast enough to be deployed on inexpensive edge computing devices that can be mass-produced. a
Transactions on Machine Learning and Artificial Intelligence, Aug 1, 2020
This survey performs a thorough enumeration and analysis of existing methods for data stream proc... more This survey performs a thorough enumeration and analysis of existing methods for data stream processing. It is a survey of the challenges facing streaming data. The challenges addressed are preprocessing of streaming data, detection and dealing with concept drifts in streaming data, data reduction in the face of data streams, approximate queries and blocking operations in streaming data.
Supervisory Control and Data Acquisition (SCADA) systems and Industrial Control Systems (ICS) are... more Supervisory Control and Data Acquisition (SCADA) systems and Industrial Control Systems (ICS) are critical for infrastructure operations, production processes, automation systems, and other automated control systems. SCADA systems are vulnerable to cyber physical attacks from many vectors. The attack surface varies widely. The problem is complicated by the inherent focus on performance, reliability, and safety rather than cybersecurity. The Secure SCADA Framework establishes a security engineered approach evolving SCADA and ICS systems towards a more cybersecure posture. The encapsulating and integrating concepts of the Secure SCADA Framework require development and analysis of implementations achieving the eight framework goals. This work provides a solution for research, design, development, and evaluation of components of a Secure SCADA System. A Security Engineered SCADA Laboratory requires a test bed with which engineering and science designs can be developed. This paper presents a test bed which can model and collect data on simulated attacks, analyze data to evaluate performance, and provide the foundations for a Security Engineered SCADA Laboratory.
Journal of Modern Applied Statistical Methods, May 1, 2005
Provided in this article is an algorithm and code for computing exact critical values (or percent... more Provided in this article is an algorithm and code for computing exact critical values (or percentiles) for Friedman's nonparametric rank test for k related treatment populations using Visual Basic (VB.NET). This program has the ability to calculate critical values for any number of treatment populations ( k ) and block sizes ( ) b at any significance level ( ) α . We developed an exact critical value table for 2(1)5 k = and 2(1)15 b = . This table will be useful to practitioners since it is not available in standard nonparametric statistics texts. The program can also be used to compute any other critical values.
The Random Forest (RF) algorithm, originally proposed by Breiman et al. (1), is a widely used mac... more The Random Forest (RF) algorithm, originally proposed by Breiman et al. (1), is a widely used machine learning algorithm that gains its merit from its fast learning speed as well as high classification accuracy. However, despite its widespread use, the different mechanisms at work in Breiman's RF are not yet fully understood, and there is still ongoing research on several aspects of optimizing the RF algorithm, especially in the big data environment. To optimize the RF algorithm, this work builds new ensembles that optimize the random portions of the RF algorithm using genetic algorithms, yielding Random Genetic Forests (RGF), Negatively Correlated RGF (NC-RGF), and Preemptive RGF (PFS-RGF). These ensembles are compared with Breiman's classic RF algorithm in Hadoop's big data framework using Spark on a large, high-dimensional network intrusion dataset, UNSW-NB15.
International Journal of Changes in Education, Jan 8, 2024
The goal of this research study was to investigate the effects of D.A. Kolb's active experimentat... more The goal of this research study was to investigate the effects of D.A. Kolb's active experimentation stage of experiential learning theory (ELT) on teacher professional development. This qualitative interpretive phenomenological study addressed the problem of teachers' desires for effective professional learning experiences to improve student learning and instruction in the classroom. The research questions explored the perceptions of private school teachers regarding professional development sessions, specifically in the context of active experimentation. The study aimed to investigate the advantages of promoting a professional culture of continuous learning and teamwork among teachers through various methods such as lab classrooms, peer observation, professional learning communities, and peer coaching. The study adopts a descriptive approach utilizing focus group discussion and semi-structured interviews for data collection. Purposeful sampling was used to determine the sample and criterion sampling was used to select the participants. Biographic data were collected to provide background information about the participants involved. The biographic data allowed for stratification enabling the examination of various factors connected to the aim of the study. The study focused on understanding the benefits of providing teacher support through active experimentation during professional development sessions. Results showed that although teachers did not explicitly reference ELT when discussing effective strategies during professional development sessions, they referenced the value of hands-on, active participation in realclassroom stimulations during teacher professional development. The study's findings also indicate a need for creating a consistent framework for professional development sessions that includes active learning, teacher collaboration, and continuous support in a culture anchored in trust.
Advances in systems analysis, software engineering, and high performance computing book series, 2015
In this chapter, we use MySQL Database Cluster to demonstrate and discover the capabilities of ke... more In this chapter, we use MySQL Database Cluster to demonstrate and discover the capabilities of key based database sharding and provide the implementation details to build a key based sharded database system. After the implementation section, we present some examples of datasets that were sharded using our implementation. The sharded data is then used for data mining, specifically association rule mining. We present the results (association rules) for the sharded data as well as the non-sharded data.
Object-oriented database systems began developing in the mid-80’s out of a necessity to meet the ... more Object-oriented database systems began developing in the mid-80’s out of a necessity to meet the requirements of applications beyond the data processing applications which were [are] served by relational database systems. This paper serves as an overview on the achievements of object-oriented database technology so far, and also discusses the weaknesses that have to be yet resolved by the object-oriented database community before object-oriented database technology can become as widespread as relational databases.
Journal of Information Technology Research, Mar 25, 2022
With the steady rise in the use of smartphones, specifically Android smartphones, there is an ong... more With the steady rise in the use of smartphones, specifically Android smartphones, there is an ongoing need to build strong intrusion detection systems to protect ourselves from malicious software attacks. This work focuses on a subgroup of android malware, scareware. The novelty of this work lies in being able to detect the various scareware families individually using a small number of network attributes, determined by a recursive feature elimination process based on information gain. No work has yet been done on analyzing the scareware families individually. Results of this work show that the number of bytes initially sent back and forth, packet size, amount of time between flows and flow duration are the most important attributes that would be needed to classify a scareware attack. Three classifiers, Decision Tree, Naïve Bayes, and OneR, were used for classification. The highest average classification accuracy (79.5%) was achieved by the Decision Tree classifier with a minimum of 44 attributes.
International Journal of Information and Communication Technology Education, 2016
The goal of this research was to investigate the effects of automated testing software on levels ... more The goal of this research was to investigate the effects of automated testing software on levels of student reflection and student performance. This was a self-selecting, between subjects design that examined the performance of students in introductory computer programming classes. Participants were given the option of using the Web-CAT software-testing tool to evaluate their computer code. Student self-reported levels of reflection were measured using Kember et al.'s (2000) reflective thinking survey.
International Journal on Engineering, Science and Technology
Computer Science drives innovation and discovery, and Computer Science classes in High Schools, p... more Computer Science drives innovation and discovery, and Computer Science classes in High Schools, particularly AP Computer Science Principles, have also been linked to better performance in other High School courses, but a Code.org report shows that the state of Florida, particularly the Northwest Florida region, is far behind the rest of the nation in offering Computer Science classes at the High School level. This paper discusses strategies that are can be implemented to improve High School Computer Science offerings in like-regions that have lower Computer Science offerings. The paper also discusses the challenges that come up in addressing this issue and presents potential avenues for addressing these challenges. One major challenge that is addressed is creating the culture of inclusive computing by getting more women and under-represented groups interested in taking Computer Science courses. Though this is a Northwest Florida case study, most of the strategies discussed in this p...
High-altitude ballooning, along with other aerospace endeavors, requires extensive preplanning an... more High-altitude ballooning, along with other aerospace endeavors, requires extensive preplanning and preparation forvehicle launching. In ballooning specifically, weather conditions are especially effective and driving whether or nota launch can occur. As most flights must be shaped around the flight path, both for safety and recovery reasons,it is imperative that any acceptable flight path and day may be considered. The goal of this project is to minimize,using machine learning, the complexity and manpower requirements for determining if a launch can occur.
The Random Forest (RF) algorithm, originally proposed by Breiman [7], is a widely used machine le... more The Random Forest (RF) algorithm, originally proposed by Breiman [7], is a widely used machine learning algorithm that gains its merit from its fast learning speed as well as high classification accuracy. However, despite its widespread use, the different mechanisms at work in Breiman's RF are not yet fully understood, and there is still ongoing research on several aspects of optimizing the RF algorithm, especially in the big data environment. To optimize the RF algorithm, this work builds new ensembles that optimize the random portions of the RF algorithm using genetic algorithms, yielding Random Genetic Forests (RGF), Negatively Correlated RGF (NC-RGF), and Preemptive RGF (PFS-RGF). These ensembles are compared with Breiman's classic RF algorithm in Hadoop's big data framework using Spark on a large, high-dimensional network intrusion dataset, UNSW-NB15.
The K-Nearest Neighbor! (KNN) algorithm is one of the most widely used algorithms in data mining ... more The K-Nearest Neighbor! (KNN) algorithm is one of the most widely used algorithms in data mining for classification and prediction. The algorithm has several applications: in facial detection when used with deep learning, in biometric security applications etc. The traditional KNN algorithm involves an iterative process of computing the distance between a test data point and every data point in the training dataset, and classifying the object based on the closest training sample. This method first selects K nearest training data points for classifying a test data point and then predicts the test sample's class based on the majority class among those neighbors. If both the train and test datasets are large, this conventional form can be considered computationally expensive. Reduction of the massive calculation that is required to predict a data vector was our main goal, and with this intention, the training dataset was split into several buckets. The KNN algorithm was then performed inside a bucket, instead of iterating over the whole training dataset. We used the Jaccard Coefficient to determine the degree of similarity of a data vector with some arbitrarily defined data points P and placed similar data points in the same bucket. This was the core functionality of our hash function. The hash function determines the bucket number where the similar data vectors will be placed. Unlike the standard hashing algorithm, our approach of hashing was to maximize the probability of the hash collision to preserve the locality sensitiveness. Both the conventional and proposed methods were implemented in Hadoop's MapReduce framework. Hadoop gives us an architecture for handling large datasets on a computer cluster in a distributed manner and gives us massive scalability. The use of the locality sensitive hashing in KNN in Hadoop's MapReduce environment took less time than conventional KNN to classify a new data object.
Vol. 19 No. 5 MAY 2021 International Journal of Computer Science and Information Security, (IJCSIS), 2021
The focus on this work is on classifying phishing emails using deep neural networks. Since phishi... more The focus on this work is on classifying phishing emails using deep neural networks. Since phishing emails have no specific characteristic, they are difficult to detect and classify, and little research has been done on the detection of phishing emails. In this work, two deep neural networks, Long Short Term Memory (LSTM), a form of Recurrent Neural Networks (RNN), and Convolutional Neural Networks (CNN), were compared and used for classification of phishing emails. RNN is the most used neural network for text classification. CNNs have also shown to be effective in text classification. In addition to tuning hyperparameters, different activation functions and optimizers are used for comparing the performance of CNN and LSTM on the basis of accuracy and the ROC-score. LSTM achieved a higher accuracy than CNN, and overall the Adam Optimizer performed better than the SGD optimizer. The best parameters for higher accuracy and ROC-score are also presented.
Uploads
Papers by Sikha Bagui