Model-Based Signal Processing & Probabilistic Inference
For projects where measurement data, statistical uncertainty, and domain knowledge need to be combined into robust analyses, interpretable models, and reproducible workflows.
Dipl.-Ing. Dr.techn. Thomas Suppan · TU Graz · Award of Excellence (Austria)
When TS-Analytics can help
TS-Analytics supports projects where complex data need to be processed, modeled, interpreted, and translated into reproducible analytical results.
Complex measurement data
You work with measurement or time-series data that require filtering, preprocessing, feature extraction, spectral analysis, time-series analysis, or structured signal evaluation.
Unknown parameters or system states
You need mathematical or probabilistic models to estimate parameters, infer system states, solve inverse problems, or describe patterns and dependencies in complex data.
Statistical evaluation and uncertainty
You need statistical analysis with uncertainty quantification, regression models, method comparison, performance metrics, or transparent evaluation of variability and model results.
Manual or recurring analysis tasks
You want to automate recurring data processing, analysis, visualization, or reporting tasks using reproducible workflows in Python, MATLAB, or R.
Results for reports or publications
You need publication-ready analyses, figures, method descriptions, technical documentation, or structured summaries for scientific and technical projects.
Services and Expertise
TS-Analytics provides specialized expertise across five connected areas of data processing, statistical modeling, signal analysis, and reproducible computational workflows.
Signal Processing
Development and implementation of signal processing methods for extracting, characterizing, and interpreting information from measurement, and time-series data.
Modeling & Inference
Model-based and probabilistic methods for parameter estimation, system-state inference, inverse problems, and interpretable model-based analysis of complex data.
Statistical Analysis
Classical and Bayesian evaluation of technical, scientific, and clinical datasets with transparent assumptions, quantitative results, and uncertainty-aware interpretation.
Automated Workflows
Design and implementation of reproducible computational workflows, automated analysis pipelines, and project-specific tools in Python, MATLAB, or R.
Scientific Support
Analytical support for scientific and technical projects, including result presentation, method descriptions, documentation, and publication-ready outputs.
View all services →
Typical Deliverables
Project outputs are designed to be reproducible, documented, and ready for review, reporting, or further development.
Analysis results and reports
Statistical reports, quantitative results, uncertainty evaluation, method comparison, performance assessment, and structured result interpretation.
Methods, models, and algorithms
Signal processing methods, probabilistic models, parameter estimates, inference algorithms, and project-specific analysis methods.
Reproducible workflows and documentation
Python, MATLAB, or R scripts, automated pipelines, reporting workflows, publication-ready figures, method descriptions, and technical documentation.
Approach
TS-Analytics follows a structured workflow from understanding the data and objective to implemented analysis methods, reproducible results, and documented outputs.
1. Understand the data and objective
Clarify the available data, measurement context, research question, analysis objective, assumptions, and required outputs.
2. Development of the analytical workflow
Select, adapt, or implement suitable signal processing, statistical modeling, inference, or automation methods for the project.
3. Generate reproducible results
Apply the workflow with transparent processing steps, documented assumptions, uncertainty evaluation, and reviewable analytical outputs.
4. Delivery of documented outputs
Provide scripts, reports, figures, models, workflows, or technical documentation that can be reviewed, repeated, and extended.
1. Understand the data and objective
Clarify the available data, measurement context, research question, analysis objective, assumptions, and required outputs.
2. Development of the analytical workflow
Select, adapt, or implement suitable signal processing, statistical modeling, inference, or automation methods for the project.
3. Generate reproducible results
Apply the workflow with transparent processing steps, documented assumptions, uncertainty evaluation, and reviewable analytical outputs.
4. Delivery of documented outputs
Provide scripts, reports, figures, models, workflows, or technical documentation that can be reviewed, repeated, and extended.
Selected Work
Selected work from academic research and scientific collaborations at TU Graz and Medical University of Graz, demonstrating the methodological foundation behind TS-Analytics.
Inverse Capacitive Flow Metering in Industrial Pneumatic Conveying Processes
Academic research · TU Graz
Model-based signal processing and inverse problem methods for electrical capacitance tomography in industrial pneumatic conveying and flow measurement applications.
Signal Processing · Inverse Problems · Measurement Data · Industrial Flow Measurement
Bayesian Meta-Analysis of NIRS-Guided Neonatal Resuscitation
Clinical research · Medical University of Graz
Implementation and application of a Bayesian meta-analysis to estimate treatment benefit probabilities for clinical neonatal outcomes.
Bayesian Meta-Analysis · Individual Patient Data · Clinical Outcomes · Probabilistic Inference
Bayesian Estimation of System States from High-Dimensional Data
Academic research · TU Graz
Probabilistic modeling and sequential inference methods for estimating discrete system states from complex, high-dimensional measurements using dimensionality reduction and a hidden Markov model.
Bayesian Inference · State Estimation · Probabilistic Models · High-Dimensional Data
Automated Quantitative Analysis of Lung Ultrasound Video Loops
Clinical research · Medical University of Graz
Implementation of a MATLAB-based analysis framework for ROI-based quantitative evaluation of neonatal lung ultrasound video loops, including grayscale and heterogeneity metrics.
MATLAB · Video Analysis · Quantitative Ultrasound · Automated Workflows
Inverse Capacitive Flow Metering in Industrial Pneumatic Conveying Processes
Academic research · TU Graz
Model-based signal processing and inverse problem methods for electrical capacitance tomography in industrial pneumatic conveying and flow measurement applications.
Signal Processing · Inverse Problems · Measurement Data · Industrial Flow Measurement
Bayesian Meta-Analysis of NIRS-Guided Neonatal Resuscitation
Clinical research · Medical University of Graz
Implementation and application of a Bayesian meta-analysis to estimate treatment benefit probabilities for clinical neonatal outcomes.
Bayesian Meta-Analysis · Individual Patient Data · Clinical Outcomes · Probabilistic Inference
Bayesian Estimation of System States from High-Dimensional Data
Academic research · TU Graz
Probabilistic modeling and sequential inference methods for estimating discrete system states from complex, high-dimensional measurements using dimensionality reduction and a hidden Markov model.
Bayesian Inference · State Estimation · Probabilistic Models · High-Dimensional Data
Automated Quantitative Analysis of Lung Ultrasound Video Loops
Clinical research · Medical University of Graz
Implementation of a MATLAB-based analysis framework for ROI-based quantitative evaluation of neonatal lung ultrasound video loops, including grayscale and heterogeneity metrics.
MATLAB · Video Analysis · Quantitative Ultrasound · Automated Workflows
View Publications & Projects →
Scientific and engineering background
TS-Analytics is led by DI Dr.techn. Thomas Suppan, an electrical engineer and applied statistician with a background in statistical signal processing, measurement systems, probabilistic modeling, and clinical research data analysis. The expertise combines engineering understanding, statistical rigor, and reproducible computational implementation — linking complex measurement data, models, uncertainty, and analytical workflows.
Statistical Signal Processing · Measurement Systems · Probabilistic Modeling · Inverse Problems · Clinical Research Data Analysis · Reproducible Workflows
More about TS-Analytics →
Get in Touch
Working on a technically or scientifically challenging project involving complex data, signals, models, or uncertainty?
Get in touch to discuss how TS-Analytics can support your analysis, modeling, or workflow development needs.
Free initial call · NDA possible · Remote collaboration across AT/DE/EU