Insights: Software that extracts knowledge from data. Model. Predict. Simulate. Discover. Associate. Gain new Knowledge and Insights from noisy data. Easily and Reliably. Ultra-fast, parallel, self-organizing, high-dimensional modeling of real-world processes and complex systems.
Insights: Self-organizing Modeling. The next generation of knowledge mining.
Today, Data Mining is widely recognized as the process of collecting and analyzing amounts of personal data of customers and people for different purposes by an increasing number of companies, governments and public services. However, Data Mining, and especially Knowledge Mining, is much more than this and can be much more useful in various problem fields. Humans have for centuries been seeking proxies for real processes. A substitute that can generate reliable information about a real system and its behaviour is called a model and they form the basis for any decision. It is worth building models to aid decision making, because models make it possible to: Identify the relationships between cause and effect. This leads to a deeper understanding of the problem at hand by deriving an analytical relationship between them, predict the respective objects can expect over a finite future time span, but also to experiment with models. Exactly the ability to make predictions about the future forms the core of intelligence at al, simulat the objects' behaviour by experiment with models, and thus answer what-if questions essential to decision-making, control the objects by finding suitable means to effect the objects and enforce a specific behaviour.
Don't care about dimensionality of your data sets any longer. High-dimensional modeling in Insights scales not only to data sets of very small to large number of samples but also from small to very large number of potential input variables. You simply self-organize models from data directly. This applies even for so-called under-determined modeling tasks, i.e., when the number of data samples is smaller than the number of input variables. Insights ensures that the appropriate modeling and noise filtering algorithm is used to get reliable and compact predictive model composits.
There are many cases in practice where it is impossible to create analytical models using classical theoretical systems analysis since there is incomplete knowledge of the processes involved. Environmental, medical and socio-economic systems are but three examples. KnowledgeMiner Software is supporting the FuturICT project, a major international public effort on modeling, understanding, and simulating our complex, global socio-eco-economic world. There are a lot of complex problems, which do need decision-making, but the means - the models - for understanding, predicting, simulating, and where possible controlling such systems are simply missing increasingly. A more powerful and easy-to-use tool that fills this knowledge gap is self-organizing modeling. Obtaining a model from data is easy. Obtaining a model from data that reliably reflects the underlying relationship in the data with some certainty is hard work. This is especially true for noisy, disturbed data. Noisy data are everywhere so you most probably will use them. To systematically avoid overfitting - that is, when the model fits to random cases (noise) and therefore can only have poor predictive power, which makes it useless -, to get optimized transfer functions in Active Neurons, and to self-organize robust optimal complex models with optimal predictive power, Insights employs original concepts of model validation at different levels of the self-organizing modeling process. Together with our Live Prediction Validation, application of data mining models is now more reliable, stable, and valuable than ever before. Until today, in data mining, model validation ends where modeling finishes. The model is then applied in real processes for prediction, classification, recommendation, diagnosis as if it would work stable and correct under all circumstances. But by far, it actually does not. Data mining models are always built on a finite set of data with certain properties. Therefore, like any model, they reflect the underlying object or process incomplete and approximately, only. Their validity is limited in space and/or time which defines their actual applicability domain. If a model works in its applicability domain or not only depends on the input data given to the model at runtime to predict an outcome. If a model works outside its domain it gets instable with irregular and presumably false prediction. How can a prediction respectively model be used with good conscience if there is no indication if it is applicable at all to the given inputs? The situation that a model works outside its applicability domain is not exceptional but can happen very often, especially for nonlinear models which are common in data mining.
We at KnowledgeMiner Software have taken this problem serious and we are proud that we came up with a powerful, original solution that makes model application in Insights much more trustworthy. For every single prediction you calculate, you also get the information if and to which extent the model is working in its domain. This makes further use of the prediction in your decision process safer and more robust.

Import your data from Excel right away into Insights by a single mouse click. This is the most convenient and fast way to start modeling from your data. But also the reverse way works: Export your developed models and selected input data to Excel and further use them in your favorite way. No matter if it is a single model or a model composite you also can have Insights generating related plots for you. Employ your models in Excel with no additional effort.
Insights is the new brand of our former KnowledgeMiner product.
© 2001-2013 KnowledgeMiner Software
Contact Us