BLIASoft Knowledge Discovery is a predictive analytics software based on artificial intelligence
techniques and in particular on fuzzy logic.
It enables to model, understand and optimize the most complex processes.
BLIASoft Knowledge Discovery has for main objectives:
• To extract predictive models.
• To facilitate the understanding of the phenomena by representing their functioning with simple and readable rules.
• To reveal the hidden interactions in order to release valuable trends and to innovate.
• To capitalize the knowledge by transforming abundant and heterogeneous data into simple and compact models.
• To improve the quality of the decisions by optimizing the choices in the environments where must be simultaneously numerous criteria and variables.
• To give value to existing and unexploited data.
BLIASoft Knowledge Discovery is particularly easy to use (friendly interface, automatic parameter setting) and does not require skills in data processing.
Transport, health, energy, finance, insurance, distribution, tourism, administrations... BLIASoft Knowledge Discovery is addressed without limitations to all the sectors as soon as it is a question of understanding, of foreseeing and of optimizing complex processes by basing only on the experience and the acquired data.
BLIASoft Knowledge Discovery concerns numerous activities of the company, such as for example:
• the R&D for the reduction of the costs and the periods of development of the new products,
• the Production in order to optimize processes,
• the Quality Department for the diagnosis of the breakdowns,
• the Marketing in order to better know the behaviour of the customers and to discover new markets,
• the Directions for a decision-making aid in the strategic choices...
An approach adapted to the data of the real world
• Imperfect data (incomplete, uncertain, contradictory)
• Heterogeneous data (qualitative and quantitative)
• From very small to very large databases, not uniformly distributed
An economic approach
An approach facilitating the understanding of the phenomena
• Understandable interpretation by rules written in natural language
• No particular skill in data processing is needed to use the software and analyse the results
Cases to be privileged
Although applicable with all the types of data, BLIASoft Knowledge Discovery takes its entire dimension in the following situations:
• For the analysis of the processes and the complex phenomena whose modeling is not available or too expensive.
• When the quality and the reliability of the information are not guaranteed (missing, badly distributed, uncertain data).
• For the domains involving human factors.
• With information already available, not envisaged originally for a treatment.
• Generally, when the traditional approaches of data processing prove to be ineffective.
It is possible to treat large-sized databases. On the contrary, our method is particularly effective with data available in very small quantity, i.e. when it is necessary to maximize the knowledge starting from a minimum of information. It is an important advantage as the cost of the collection or the creation of the data is often high.
Examples of possible realizations
Optimization of industrial processes
Understand and control the complex processes governed by a large number of parameters and for which there are no satisfactory tools of modeling. Determine the really influential parameters and their interactions by exploiting concrete cases from production.
Reduction of development costs
By taking advantage of in-service experience, expertise and skills of employees or trial results from previous or similar products, the duration and cost of conception and validation phases can be reduced.
Detection of dysfunctions
Exploitation of data relative to industrial processes or behaviours in order to clarify the situations having led to abnormalities or frauds. Extract models allowing the prediction of future deviations.
Behaviour of customers
Analyze and model the behaviour of the customers in order to optimize advertising and marketing campaigns thanks to the treatment of the information resulting from direct observations, inquiries or sales statistics.
Forecast of the natural phenomena
Exploitation of reports and observation with the aim of modeling and predicting at lower cost phenomena usually requiring important calculations (example: weather forecasts).
BLIASoft Knowledge Discovery provides graphical tools for 2D, 3D and 4D visualization in order to facilitate the analysis of the solutions (post-processing).
Based on Microsoft® DirectX® technology, the 3D graphical engine offers a powerful and easy-to-use interface with a lot of the functionalities (zoom, rotations, transparency, …).
Practical cases (benchmarks)
This rubric proposes to illustrate the functioning of BLIASoft Knowledge Discovery by the study of some practical cases of datasets.
Some of these data sets come from the UCI Machine Learning Repository (http://archive.ics.uci.edu/ml/index.html), Irvine, CA, University of California, School of Information and Computer Science.