Search engines
For creating search engines (basic online version, we developed an innovative and rigorous methodology (principles of operation) based on measuring how far items are away form being a Pareto optimum. Thus we call the central quantitative concept of the methodology and of our custom-made search engines k-Pareto optimality.
Theoretical properties, examples and practical applications of k-Pareto optimality based sorting have have been published at the AISTATS 2022 conference by Jean Ruppert (Mathematics and Computing S.à.r.l.) and Marharyta Aleksandrova and Thomas Engel (University of Luxembourg).
Advantages of sorting by k-Pareto optimality
Our search engines sort data by k-Pareto optimality and guarantee
- maximisation of choice and diversity of the results,
- respect of the user's preferences,
- respect of economic common sense by giving a higher value to rarer characteristics,
- scores and sub-scores that are easy to understand.
Offer
The source code of the generic search engine is available against a license fee. Customisation and integration into your production environment will be provided as a service.
Other uses of the k-Pareto optimality based methodology
The methodology based on sorting by k-Pareto optimality is flexible and allows for intuitive parameterisation via two mathematical structures.
- A measure indicates the importance, size or likelihood of objects,
- a transitive relation indicates preferences, dependence or causality.
Besides search engines the underlying methodology can therefore also be used for
- duplicate detection tools,
- outlier detection tools,
- statistical tests,
- statistical indicators,
- risk management and project management tools.