Probabilistic Knowledge Processing
Knowledge based systems started with purely deterministic rule processing in the early 1970's. They are nowadays able to manage uncertain, subjective and vague knowledge. One research area considers probabilistic facts and rules suitable for communication between user and system, since probabilities are particularly able to quantify given dependencies. In this context probabilities have either be estimated by an expert or can be calculated using statistical data.
The expert system shell SPIRIT offers a good opportunity to represent and process uncertain data. Therefor frequency distributions can be used as well as heuristic rules, expert’s knowledge as well as subjective judgement. Furthermore, SPIRIT is able to learn from data and examples. The evaluation of given information and generation of a probabilistic knowledge base render it possible to generate not noted associations. It's quite remarkable that the system is able to derive implications within the meaning of predicate logic based on a common probability distribution.
SPIRIT is a sophisticated expert system shell allowing an easy user communication. The acronym stands for Symmetrical Probabilistic Intensional Reasoning in Inference Networks in Transition. SPIRIT-Shell is a program for generating, processing and applying probabilistic SPIRIT-knowledge-bases. The construction of a knowledge base is performed in an optimal way allowing partial information about the domain, too. SPIRIT was developed in the platform-independent programming language Java and a trial version is free at your disposal.
SPIRIT-Shell is a software-package developed at our university. We are trying to add helpful features and so expanding the program. Please don't hesitate to send comments and suggestions for improvement to Prof. W. Rödder; he will transmit them to the developers team immediately.
You can download SPIRIT-Shell as Application and examine it for yourself free of charge. To run the program, a Java runtime environment is needed. If such a system is still not installed on your computer, we recommend to use the JRE [ext] from Sun, which is also free available.
Please go to our download page [ext] now!
Project Management and Research
Specialized Research on Theoretical Topics
- Dr. Carl Heinz Meyer • Basic Algorithms
- Dr. Friedhelm Kulmann • Information Theory and Related Topics
- Dr. Andreas Dellnitz • Social Network Analysis
- Prof. Dr. Elmar Reucher • Utility and Decision Analysis
- Prof. Dr. André Ahuja • Riskmanagement
- Longgui Xu (Prof. VR China) • Consistency Analysis
- Dr. Holger Knublauch [ext] • System Programming
- Dr. Carl Heinz Meyer • System Conception
- Dr. Friedhelm Kulmann [mailto]
Social Network Analysis
The analysis of Social Networks widely is based on their respective graphical structures. Centrality measures of actors usually consider their position in a graph. Sometimes graphs are an insufficient medium to represent social structures. The research group proposed a new framework with the purpose to analyze social fabric correctly: information theory. A set of conditionals is used to form the knowledge base about the network’s structure. Once the knowledge base is adapted under maximum entropy or minimum cross-entropy a new analysis tool is available. The construction of knowledge bases and their analyses are realized in the expert system shell SPIRIT.
Assessment of Quotes in Plant Business
Until conclusion of a contract for the purchase of a large-scale plant the potential buyer obtains offers from different suppliers. For the manufacturer the preparation of a tender is connected with substantial costs. But there is no guarantee to win the tender – only ten percent of these offers are successful. The costs of project preparation in plant engineering amounts to five percent of the order value, and thereby in the high millions of euros. So there is a special interest to rate the chances winning the tender. An assessment of quotes can be performed by the expert-system shell SPIRIT in two steps. First a check of the basic conditions occurs in a preselection, before then on the base of an extensive knowledge base chances and risks are forecast for submitting and keeping open a tender respectively. Within the scope of a project with a big German plant engineering the application of such a system was tested. For different projects the analysis by SPIRIT was compared with alternative assessment methods.
- W. Rödder; A. Dellnitz; F. Kulmann; S. Litzinger; E. Reucher: Bipartite Structures in Social Networks – traditional versus entropy-driven analyses, Entropy, 21:3 (2019) 277. DOI: 10.3390/e21030277
- Rödder, W.; Dellnitz, A.; Kulmann, F.; Litzinger, S.; Reucher, E.: Bipartite Strukturen in Sozialen Netzen -- klassische versus MaxEnt-Analysen, Diskussionsbeitrag der Fakultät für Wirtschaftswissenschaft, FernUniversität in Hagen, 503 (2017).
- Brenner, D.; Dellnitz, A.; Kulmann, F.; Rödder, W.: Compressing strongly connected subgroups in social networks -- an entropy-based approach, The Journal of Mathematical Sociology, 41 (2017) 84-103. DOI: 10.1080/0022250X.2017.1284070.
- Dellnitz, A.; Litzinger, S.; Rödder, W.: Identifizierung von führenden Köpfen in terroristischen Netzwerken - ein entropiebasiertes Verfahren - , Diskussionsbeitrag der Fakultät für Wirtschaftswissenschaft, FernUniversität in Hagen, 507 (2017).
- Rödder, W.; Kulmann, F.; Dellnitz, A.: A New Rationality in Network Analysis – Status of Actors in a Conditional-logical Framework, In: Beierle, C.; Brewka, G.; Thimm, M. (editors), Computational Models of Rationality, Tributes Vol.20 (College Publications, 2016) 348-364.
- Rödder, W.; Brenner, D.; Kulmann, F.: Entropy based evaluation of net structures – deployed in Social Network Analysis, Expert Systems with Applications, 41-17 (2014) 7968–7979. DOI: 10.1016/j.eswa.2014.06.049
- Brenner, D.; Rödder, W.; Kulmann, F.: Superknoten in Sozialen Netzwerken – eine entropieoptimale Analyse, Diskussionsbeitrag der Fakultät für Wirtschaftswissenschaft, FernUniversität in Hagen, 485 (2013).
- Rödder, W.; Brenner, D.; Kulmann, F.: Entropiebasierte Analyse von Interaktionen in Sozialen Netzwerken, Diskussionsbeitrag der Fakultät für Wirtschaftswissenschaft, FernUniversität in Hagen, 481 (2012).
- Rödder, W.; Reucher, E.; Kulmann, F.: Where We Stand at Probablistic Reasoning, Proc. 6th Int. Conf. on Informatics in Control, Automation, and Robotics ICINCO 2009, July 2-5, Mailand, Italien (2009) 394-397.
- Reucher, E.; Kulmann, F.: Probabilistic Knowledge Processing and Remaining Uncertainty , Proc. 20th Int. FLAIRS Conf. - FLAIRS-20, May 7-9, Key West, Florida (2007) 122-127.
- Rödder, W.; Reucher, E.; Kulmann, F.: Features of the Expert-System-Shell SPIRIT, Logic Journal of the IGPL, 14-3 (2006) 483-500.
- Rödder, W.; Kulmann, F.: Recall and Reasoning - an Information Theoretical Model of Cognitive Processes , Information Sciences, 176-17 (2006) 2439-2466.
- Kern-Isberner, G.; Rödder, W.; Kulmann, F.: Conditionals, Information, and Inference CII'02, International Workshop, Lecture Notes in Artificial Intelligence, Vol. 3301 (Springer, Berlin/Heidelberg, 2005).
- Kulmann, F.: Wissen und Information in konditionalen Modellen - Zur Entscheidungsvorbereitung im Anfrage- und Auftragsmanagement, (Deutscher Universitäts-Verlag, Wiesbaden, 2002).
- Kulmann, F.; Reucher, E.: Computergestützte Bonitätsprüfung bei Banken und Handel , DBW - Die Betriebswirtschaft , 60 (2000) 113-122.