Steven Walczak
University of South Florida, School of Information, Department Member
This study investigates the use of artificial neural networks (ANNs) to classify reasons for medication nonadherence. A survey method is used to collect individual reasons for nonadherence to treatment plans. Seven reasons for... more
This study investigates the use of artificial neural networks (ANNs) to classify reasons for medication nonadherence. A survey method is used to collect individual reasons for nonadherence to treatment plans. Seven reasons for nonadherence are identified from the survey. ANNs using backpropagation learning are trained and validated to produce a nonadherence classification model. Most patients identified multiple reasons for nonadherence. The ANN models were able to accurately predict almost 63 percent of the reasons identified for each patient. After removal of two highly common nonadherence reasons, new ANN models are able to identify 73 percent of the remaining nonadherence reasons. ANN models of nonadherence are validated as a reliable medical informatics tool for assisting healthcare providers in identifying the most likely reasons for treatment nonadherence. Physicians may use the identified nonadherence reasons to help overcome the causes of nonadherence for each patient.
Research Interests: Information Systems, Computer Science, Health Informatics, Nigeria, Medicine, and 13 moreDrug, Library and Information Studies, Pharmaceutical, Ehr, Prescription, Medication, Backpropagation, Artificial Intelligence and Robotics, Medicine and Health Sciences, Pharmacy and Pharmaceutical Sciences, Artificial Neural Network, Nonadherence, and Medical prescription
Attention deficit hyperactivity disorder (ADHD) is a common and chronic mental health disorder found around the world. The worldwide rate for diagnosis of ADHD is just over 5% of the population and appears to be consistent for most of the... more
Attention deficit hyperactivity disorder (ADHD) is a common and chronic mental health disorder found around the world. The worldwide rate for diagnosis of ADHD is just over 5% of the population and appears to be consistent for most of the world [1]. Within the United States, ADHD is the most commonly occurring childhood mental health disorder and incidence rates as high as 11% have been reported [2]. These numbers may well be inaccurate as ADHD can be misdiagnosed, depending on the diagnostic methodology utilized [3,4]. Prior research has indicated that the prevalence of ADHD is rising significantly in the United States, by almost 22% over a four year period [5]. ADHD occurs across all socioeconomic, cultural, and racial backgrounds [6] and is therefore a concern for all pediatric patients.
Neural networks are a machine learning method that excel in solving classification and forecasting problems. They have also been shown to be a useful tool for working with big data oriented environments such as law enforcement. This... more
Neural networks are a machine learning method that excel in solving classification and forecasting problems. They have also been shown to be a useful tool for working with big data oriented environments such as law enforcement. This article reviews and examines existing research on the utilization of neural networks for forecasting crime and other police decision making problem solving. Neural network models to predict specific types of crime using location and time information and to predict a crime’s location when given the crime and time of day are developed to demonstrate the application of neural networks to police decision making. The neural network crime prediction models utilize geo-spatiality to provide immediate information on crimes to enhance law enforcement decision making. The neural network models are able to predict the type of crime being committed 16.4% of the time for 27 different types of crime or 27.1% of the time when similar crimes are grouped into seven categ...
Research Interests:
First-time leaders may find themselves thrust into very stressful situations for their teams and organizations at large. First-time leaders in corporations, the classroom, sports, the military, and politics should understand how stress... more
First-time leaders may find themselves thrust into very stressful situations for their teams and organizations at large. First-time leaders in corporations, the classroom, sports, the military, and politics should understand how stress changes the way followers perceive their leader and the ideal traits for a leader through changing leadership prototype schemas. Implicit leadership theories, social information processing, and cognitive psychology suggest that stress can influence the activation of schema. Changing leadership prototype schemas of followers may affect subsequent productivity and efficiency. This chapter examines if leadership prototype schemas change under stress and recommends ways first-time leaders can respond to these changing schemas, including how female first-time leaders who are often initially perceived as more sensitive leaders can utilize changing perceptions and ideal leader prototypes under stressful conditions.
Cyberbullying is a growing and significant problem in today’s workplace. Existing automated cyberbullying detection solutions rely on machine learning and deep learning techniques. It is proven that the deep learning-based approaches... more
Cyberbullying is a growing and significant problem in today’s workplace. Existing automated cyberbullying detection solutions rely on machine learning and deep learning techniques. It is proven that the deep learning-based approaches produce better accuracy for text-based classification than other existing approaches. A novel decentralized deep learning approach called MaLang is developed to detect abusive textual content. MaLang is deployed at two levels in a network: (1) the System Level and (2) the Cloud Level, to tackle the usage of toxic or abusive content on any messaging application within a company’s networks. The system-level module consists of a simple deep learning model called CASE that reads the user’s messaging data and classifies them into abusive and non-abusive categories, without sending any raw or readable data to the cloud. Identified abusive messages are sent to the cloud module with a unique identifier to keep user profiles hidden. The cloud module, called KIPP...
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Background Best practice “bundles” have been developed to lower the occurrence rate of surgical site infections (SSI’s). We developed artificial neural network (ANN) models to predict SSI occurrence based on prophylactic antibiotic... more
Background Best practice “bundles” have been developed to lower the occurrence rate of surgical site infections (SSI’s). We developed artificial neural network (ANN) models to predict SSI occurrence based on prophylactic antibiotic compliance. Methods Using the American College of Surgeons National Quality Improvement Program (ACS-NSQIP) Tampa General Hospital patient dataset for a six-month period, 780 surgical procedures were reviewed for compliance with SSI guidelines for antibiotic type and timing. SSI rates were determined for patients in the compliant and non-compliant groups. ANN training and validation models were developed to include the variables of age, sex, steroid use, bleeding disorders, transfusion, white blood cell count, hematocrit level, platelet count, wound class, ASA class, and surgical antimicrobial prophylaxis (SAP) bundle compliance. Results Overall compliance to recommended antibiotic type and timing was 92.0%. Antibiotic bundle compliance had a lower incide...
Research Interests:
Intraductal papillary mucinous neoplasms (IPMN) are a type of mucinous pancreatic cyst. IPMN have been shown to be pre-malignant precursors to pancreatic cancer, which has an extremely high mortality rate with average survival less than 1... more
Intraductal papillary mucinous neoplasms (IPMN) are a type of mucinous pancreatic cyst. IPMN have been shown to be pre-malignant precursors to pancreatic cancer, which has an extremely high mortality rate with average survival less than 1 year. The purpose of this analysis is to utilize methodological triangulation using artificial neural networks and regression to examine the impact and effectiveness of a collection of variables believed to be predictive of malignant IPMN pathology. Results indicate that the triangulation is effective in both finding a new predictive variable and possibly reducing the number of variables needed for predicting if an IPMN is malignant or benign.
Research Interests:
Artificial neural networks are a machine learning method ideal for solving classification and prediction problems using Big Data. Online social networks and virtual communities provide a plethora of data. Artificial neural networks have... more
Artificial neural networks are a machine learning method ideal for solving classification and prediction problems using Big Data. Online social networks and virtual communities provide a plethora of data. Artificial neural networks have been used to determine the emotional meaning of virtual community posts, determine age and sex of users, classify types of messages, and make recommendations for additional content. This article reviews and examines the utilization of artificial neural networks in online social network and virtual community research. An artificial neural network to predict the maintenance of online social network “friends” is developed to demonstrate the applicability of artificial neural networks for virtual community research.
Research Interests:
Abstract Touchscreen tablet technology is being widely adopted in primary and secondary schools throughout the world. Current research largely explores how to use this technology to teach reading and writing, mathematics, and to a lesser... more
Abstract Touchscreen tablet technology is being widely adopted in primary and secondary schools throughout the world. Current research largely explores how to use this technology to teach reading and writing, mathematics, and to a lesser extent science. However a research gap exists in exploring tablet technology to teach geography. The research in this article examines if any differences in learning outcomes exist between a more traditional teaching method and one that is centered on using touchscreen tablet technology when teaching USA states’ shapes and locations to second-graders. The results indicate that there is no statistically significant difference between the two teaching methods, but that combining the two methods may lead to significant improvements in learning outcomes.
Research Interests:
Clinical decision support systems are meant to improve the quality of decision-making in healthcare. Artificial intelligence is the science of creating intelligent systems that solve complex problems at the level of or better than human... more
Clinical decision support systems are meant to improve the quality of decision-making in healthcare. Artificial intelligence is the science of creating intelligent systems that solve complex problems at the level of or better than human experts. Combining artificial intelligence methods into clinical decision support will enable the utilization of large quantities of data to produce relevant decision-making information to practitioners. This article examines various artificial intelligence methodologies and shows how they may be incorporated into clinical decision-making systems. A framework for describing artificial intelligence applications in clinical decision support systems is presented.
Research Interests:
Research Interests:
Artificial intelligence (AI) in general and artificial neural networks (ANN) in particular provide a tremendous amount of knowledge to improve managerial decision making. Additionally, these same ANN and AI techniques also serve as... more
Artificial intelligence (AI) in general and artificial neural networks (ANN) in particular provide a tremendous amount of knowledge to improve managerial decision making. Additionally, these same ANN and AI techniques also serve as knowledge repositories and distribution schema for organizations that facilitate managerial leadership responsibilities. This article examines how various ANN and other AI applications may be adapted to facilitate managerial leadership, improve manager performance and in some cases perform management activities. Further research that classifies leadership styles and the desired qualities of leaders is reviewed.
Current and future information systems require a better understanding of the interactions between users and systems in order to improve system use, and ultimately, success. The use of personas as design tools is becoming more widespread... more
Current and future information systems require a better understanding of the interactions between users and systems in order to improve system use, and ultimately, success. The use of personas as design tools is becoming more widespread as academicians and practitioners discover its benefits. This paper presents an empirical study comparing the performance of existing qualitative and quantitative clustering techniques at the task of identifying personas and grouping system users into those personas. A method based on Factor (Principal Component) Analysis outperforms two others using Latent Semantic Analysis and Multivariate Cluster Analysis.
Research Interests:
Research Interests:
The widespread use of the routine EEG in clinical practice was a major development in the treatment of patients with ill-defined spells thought to be epileptic. Not every finding on the EEG is suggestive of epilepsy, and the EEG is... more
The widespread use of the routine EEG in clinical practice was a major development in the treatment of patients with ill-defined spells thought to be epileptic. Not every finding on the EEG is suggestive of epilepsy, and the EEG is subject to over-interpretation, which may lead to misdiagnosis and incorrect treatment. Although supplemented by other procedures, the EEG remains a cost-effective and noninvasive way to diagnose spells. To enhance further the diagnostic use of the EEG, it is important to determine how strongly patterns are correlated with clinical seizures. The authors studied one EEG pattern, lateralized bursts of theta, and found the rhythmicity of the pattern to be most strongly correlated with seizures.
Research Interests:
Research Interests:
The development of multiple agent systems faces many challenges, including agent coordination and subdivision of tasks. Minsky's "The Society of Mind" [1] provides a framework for... more
The development of multiple agent systems faces many challenges, including agent coordination and subdivision of tasks. Minsky's "The Society of Mind" [1] provides a framework for addressing these multi-agent system problems. Following this framework, a society of agents is developed and applied to the domain of single player and multiplayer games. The advantage of the society approach is the efficient
Research Interests:
Communities of practice (COPs) have been around since the founding of the first social networks many millennia ago. Organizations around the world over the last two decades have leveraged COPs as tools in the knowledge management (KM)... more
Communities of practice (COPs) have been around since the founding of the first social networks many millennia ago. Organizations around the world over the last two decades have leveraged COPs as tools in the knowledge management (KM) toolkit; however, the rules of the game have changed as the ubiquity of the Internet in organizations has led to profound changes in
Research Interests:
Research Interests:
Research Interests:
Research Interests:
Neural networks have been repeatedly shown to outperform traditional statistical modeling techniques for both discriminant analysis and forecasting. While questions regarding the effects of architecture, input variable selection, learning... more
Neural networks have been repeatedly shown to outperform traditional statistical modeling techniques for both discriminant analysis and forecasting. While questions regarding the effects of architecture, input variable selection, learning algorithm, and size of training sets on the neural network model’s performance have been addressed, very little attention has been focused on distribution effects of training and out-of-sample populations on neural network performance. This article examines the effect of changing the population distribution within training sets for estimated distribution density functions, in particular for a credit risk assessment problem.
Research Interests:
Research Interests:
Opening books play a vital part in the performance of current chess programs. Use of the opening book places the chess program in a position that is already several moves into the game and can be positionally inferior. A method for... more
Opening books play a vital part in the performance of current chess programs. Use of the opening book places the chess program in a position that is already several moves into the game and can be positionally inferior. A method for acquiring opening knowledge about a specific opponent is presented. The method analyzes the historic performance of the opponent to find the opening sequences that are known to the opponent. Knowledge about the opening preferences of an opponent affords a strategic advantage to a chess program. The performance of this method is demonstrated and analyzed. Current chess programs that utilize knowledge about the opening repertoire of an opponent will be able to decrease the size of their opening books and can develop a game strategy from the start of the chess game instead of the beginning of the middle game.
Research Interests:
Medical practitioners are under ever increasing pressure to maximize patient care, while minimizing costs. One productivity area that has not previously undergone thorough investigation is the efficient utilization of time for... more
Medical practitioners are under ever increasing pressure to maximize patient care, while minimizing costs. One productivity area that has not previously undergone thorough investigation is the efficient utilization of time for intra-office communication. Medical office personnel typically need to communicate patient information and resource requests, as well as personal messages. An intra-office communication system is designed that reduces time-waste typically incurred in medical office environments. Redesigning medical offices with intra-office communication systems provides time savings of several man hours per day. The subsequent increase in time efficiency enables higher quality of patient care and larger patient loads to be managed by existing medical staff.
Research Interests: Information Systems, Communication, Organizational Culture, Telecommunications, Communication System, and 9 moreSystem Analysis and Design, Systems Analysis and Design, Medical systems, Computer User Interface Design, Public health systems and services research, Object Oriented, Equipment Design, Time and Motion Studies, and Dental offices
Research Interests: Information Systems, Communication, Organizational Culture, Telecommunications, Communication System, and 9 moreSystem Analysis and Design, Systems Analysis and Design, Medical systems, Computer User Interface Design, Public health systems and services research, Object Oriented, Equipment Design, Time and Motion Studies, and Dental offices
A hospital laboratory relational database, developed over eight years, has demonstrated significant cost savings and a substantial financial return on investment (ROI). In addition, the database has been used to measurably improve... more
A hospital laboratory relational database, developed over eight years, has demonstrated significant cost savings and a substantial financial return on investment (ROI). In addition, the database has been used to measurably improve laboratory operations and the quality of patient care.
Research Interests:
When data quantities are fixed, as they are for most financial modeling techniques at the time of model construction, a problematic issue in developing optimal models including neural network models is the selection of training and... more
When data quantities are fixed, as they are for most financial modeling techniques at the time of model construction, a problematic issue in developing optimal models including neural network models is the selection of training and validation sets within the data. A general heuristic used with time-series models and especially with neural network time-series forecasting models is that as the size of the training data set increases, then the accuracy of the time-series model also increases. However, acquiring additional model building data increases the time and financial costs of a model. This chapter investigates the effect of increasing the quantity of data used for training/building neural network forecasting models in the domain of currency exchange. Results indicate that a minimum of one to four years of data is capable of producing maximum performance neural network forecasting models, achieving forecasting accuracy of sixty percent or greater for many currency exchange rates.
Research Interests:
Based on social categorisation theory and the similarity attraction paradigm, we propose that a high degree of ethno/lingo diversity in knowledge-intensive work groups, especially short-term ones, can inhibit knowledge sharing among... more
Based on social categorisation theory and the similarity attraction paradigm, we propose that a high degree of ethno/lingo diversity in knowledge-intensive work groups, especially short-term ones, can inhibit knowledge sharing among members. Barriers identified in the knowledge sharing literature include trust, communication, and informal social interaction. A semi-quantitative approach was applied, involving a group collaborative decision making task. The results of the research task show that knowledge sharing within groups as a whole suffers as ethnic and linguistic diversity increase for short term groups. The findings provide a starting point for theory development and suggest creating more homogeneous groups for short-term tasks.
Research Interests:
The housing market is a key component of the US economy. Stability of the housing market and equity of residential property help to determine consumer confidence and their net worth. Confidence for homeowners is key to any consumer driven... more
The housing market is a key component of the US economy. Stability of the housing market and equity of residential property help to determine consumer confidence and their net worth. Confidence for homeowners is key to any consumer driven economy like that of the United States. However, a decade of low interest rate, lack of a basic credit standard, greed, and competitions among lending institutions created a housing bubble that eventually burst around the middle of 2007. This caused a severe financial meltdown in 2008. While other papers look at the meltdown from a financial market perspective, this paper will look into consumers' ability and biases towards selection of a quality mortgage. We examine various factors including educational background, risk aversion, investment self efficacy, and social position that influence consumers' ability to choose an appropriate mortgage. The results indicate that investment self efficacy has at least some impact on the quality of the mortgage decision.
Research Interests:
Research Interests:
Research Interests:
University admissions offices are flooded every year by student applicants seeking to be enrolled at the university. Depending on the particular university, twenty percent or fewer of these applicants actually become students. During... more
University admissions offices are flooded every year by student applicants seeking to be enrolled at the university. Depending on the particular university, twenty percent or fewer of these applicants actually become students. During these hard economic times for the academic community, the acquisition and retention of as many suitable applicants as possible is desirable. This paper describes a neural network system, ADMIT, which has been developed to determine the likelihood that a student applicant, if accepted, will actually attend a particular university. The ADMIT neural network enables admissions counselors to spend their time more effectively
Research Interests:
ABSTRACT
Research Interests:
... emerging more rapidly than any disease in history (Safran et al., 1996) and new pharmaceuticals are being created every day (Swerdlow ... over 100,000 medical websites (Ansani et al., 2005) and over 300,000 new medically related pages... more
... emerging more rapidly than any disease in history (Safran et al., 1996) and new pharmaceuticals are being created every day (Swerdlow ... over 100,000 medical websites (Ansani et al., 2005) and over 300,000 new medically related pages added each year (Detmer and Shortliffe ...
Research Interests:
ABSTRACT Establishing and maintaining relationships online is becoming ever more important in the expanding global knowledge economy. This study examines factors underlying the termination of relationships in the social networking... more
ABSTRACT Establishing and maintaining relationships online is becoming ever more important in the expanding global knowledge economy. This study examines factors underlying the termination of relationships in the social networking environment of Facebook and examines how this information may be used to promote virtual business relationships. The results indicate that people use a variety of reasons for terminating virtual relationships which may be broadly clustered into offline and online reasons.
Research Interests:
Artificial intelligence programs operating in competitive domains typically use brute-force search if the domain can be modeled using a search tree or alternately use nonsearch heuristics as in production rule-based expert systems. While... more
Artificial intelligence programs operating in competitive domains typically use brute-force search if the domain can be modeled using a search tree or alternately use nonsearch heuristics as in production rule-based expert systems. While brute-force techniques have recently proven to be a viable method for modeling domains with smaller search spaces, such as checkers and chess, the same techniques cannot succeed
Research Interests:
... emerging more rapidly than any disease in history (Safran et al., 1996) and new pharmaceuticals are being created every day (Swerdlow ... over 100,000 medical websites (Ansani et al., 2005) and over 300,000 new medically related pages... more
... emerging more rapidly than any disease in history (Safran et al., 1996) and new pharmaceuticals are being created every day (Swerdlow ... over 100,000 medical websites (Ansani et al., 2005) and over 300,000 new medically related pages added each year (Detmer and Shortliffe ...
Research Interests:
Patients face a multitude of diseases, trauma, and related medical problems that are difficult and costly to diagnose with respect to direct costs, including pulmonary embolism (PE). Advanced decision-making tools such as artificial... more
Patients face a multitude of diseases, trauma, and related medical problems that are difficult and costly to diagnose with respect to direct costs, including pulmonary embolism (PE). Advanced decision-making tools such as artificial neural networks (ANNs) improve diagnostic capabilities for these problematic medical conditions. The research in this chapter develops a backpropagation trained ANN diagnostic model to predict the occurrence of PE. Laboratory database values for 292 patients who were determined to be at risk for a PE, with 15% suffering a confirmed PE, are collected and used to evaluate various ANN models’ performance. Results indicate that using ANN diagnostic models enables the leveraging of knowledge gained from standard clinical laboratory tests, significantly improving both overall positive predictive and negative predictive performance.