Foresight Logic
Foresight Logic provides consulting and software development services in research, data exploration, analytics, and machine learning. In delivering these services, we build reusable software components and libraries that support development of sophisticated solutions to the knowledge-discovery and decision-making challenges of today.
Our development group draws upon many years of experience in data exploration, analytics and machine reasoning to create toolkits and frameworks that help increase the information value of data-driven decision support applications.
Statistical Analysis
Whether the data are captured through randomized control trials, naturalistic studies, quality improvement initiatives or surveys, we have the expertise to provide methodological guidance and analytic services.
Psychometrics
We design, build and validate screens and assessments using conventional test and item response theories. We work with you to clearly identify the psychometric constructs you want to measure, develop strong candidate item pools, design and execute all stages of test development. From Rasch analysis to exploratory and confirmatory factor analyses, from reliability analyses to comuterized adaptive testing, and from convergent to discriminant validity, we have the experience and the tools to build psychometrically sound tests and measures.
Software Development
We draw upon the depth and breadth of open source and commercial applications, toolkits and frameworks. In many cases coding is required to “glue” components together to form a robust and reusable workflow from knowledge discovery to decision making and feedback. When an existing solution is not readily available, sometimes the best option is to design and build your own. We draw upon a broad range of programming languages – including C#, F#, Python, SQL, R, Stata, SPSS, HTML, and Javascript to cover the full scope of analytic software development.
Machine Learning
We employ both statistical and emergent techniques to discover patterns, clusters and anomalies in data. The presenting questions and available data determine our approach. Statistical approaches include cluster analysis, factor analysis, and regression; emergent approaches include neural networks, evolutionary programming, and recursive partitioning. When the data are particularly noisy or imprecise, we augment our approach with the application of fuzzy set theory.
We strive for transparency in the recommendations, classifications and predictions that are generated. We also work to establish robust feedback pathways that enable data driven systems to continually learn from experience and improve performance over time.
DECISION SUPPORT
We employ state of the art dynamic data visualization techniques that enable decision makers to interact with real time data and seek answers to immediate business questions. We also employ techniques such as case-based reasoning to dynamically build models based on the presenting problem, make recommendations, and back those recommendations with rule-based explanations. While it is difficult for such techniques to be completely transparent, we work hard to minimize the opacity of the proverbial “black box.”
Paul Bergmann
Founder, Chief Scientist
Prior to founding Foresight Logic in 1998, Paul Bergmann served as Vice President and Chief Technology Officer of New Standards, Inc. where he pioneered the development and application of machine learning technologies to analytic and decision-making challenges in health care and finance. Over the past 20 years, Paul has provided analytic software and services to a wide array of clients in healthcare, academia, business and industry.
Paul brings an extensive background in research, software development, data analysis, data mining, and decision support to his work at Foresight Logic. His research contributions include measuring the effectiveness of behavioral health interventions, evaluating the effect of technology and clinical decision support on healthcare delivery, quantifying the impact of integrating behavioral health care with primary care, and developing and evaluating psychometric assessments. His technological contributions have included implementing RNA-Seq analytic pathways for identifying differential exon expression, refining algorithms for ground penetrating radar applications, and applying non-parametric machine learning methods to classification and prediction challenges in healthcare, target marketing and finance.
Long before data science was “a thing”, Paul was immersed in practical applications of fuzzy logic, case-based reasoning, tree-based algorithms, neural networks and information theory. His approach to any analytic challenge is rigorous and unconstrained by conventional thinking.
How Can We Help?
Let's explore how Foresight Logic can help your data speak, and help you respond.
