Insights 2.0 for OS X: Self-Organizing, Cost-Sensitive, Data Modeling

By on Jul 29, 2013 at 3:03 am

Berlin, Germany – KnowledgeMiner Software today is pleased to announce Insights 2.0 for Mac OS X, an update to its innovative application that brings conventional data mining to a new level of sophistication and applicability. The critically acclaimed application now includes purpose-driven, self-organizing, cost-sensitive modeling from imbalanced data sets. With Insights 2.0, it is now possible to develop tailored, purpose-driven, cost-optimized classification models in an easy and automated way, particularly useful in biotech, medicine, life sciences, and chemistry. Capable of computationally intensive, self-organizing modeling, the app supports multi-core, parallel processing. Insights is currently used by NASA, Boeing, MIT, Columbia University, Merck, Mobil, Notre Dame, Pfizer, Apple, and many other corporations, universities, research institutes, and individuals worldwide.

Feature Highlights:
* Brings high-performance, cost-sensitive, predictive modeling to users, with unprecedented ease of model building and deployment, and it takes full advantage of the computing power of parallel processing, available on multi-core Macs
* Hides all complex processes of knowledge extraction, model development, dimension reduction, variables selection, noise filtering, and model validation from the user
* Automatically develops models and model ensembles from data, and it generates the equation that describes the data as a mechanistic interpretation of how the model works
* Live Prediction Validation technology – for the first time, Insights instantly provides direct information about the model applicability domain vis-a-vis new data
* Checks if, and the extent to which, the developed model reflects a valid relationship, or if it just models noise, making the model useless for predictive purposes

While Insights is capable of modeling data sets of any kind, its latest update has been especially optimized for researchers affiliated with biotech firms, chemical companies, and research institutes. Many applications in biotech, medicine, life sciences, and chemistry can take advantage of this powerful tool to improve customer safety, reduce health and environmental risks of chemical products, substantially decrease research and development costs, and clearly help to minimize animal tests by using alternative, computer-based testing methods, known as QSARs.

According to Wikipedia, “Quantitative structure-activity relationship models (QSAR models) are regression or classification models used in the chemical and biological sciences and engineering. Like other regression models, QSAR regression models relate a set of ‘predictor’ variables (X) to the potency of the response variable (Y), while classification QSAR models relate the predictor variables to a categorical value of the response variable. In QSAR modeling, the predictors consist of physico-chemical properties or theoretical molecular descriptors of chemicals; the QSAR response-variable could be a biological activity of the chemicals. QSAR models first summarize a supposed relationship between chemical structures and biological activity in a data-set of chemicals. Second, QSAR models predict the activities of new chemicals.”

Insights comes with several examples of modeled data sets, such as cost-optimized models for fetal state monitoring, gene expression of tumor tissue, and QSAR models for screening and regulatory purposes. More than 15 such examples are included for illustrative purposes, taken from the fields of Climate Change, Energy, Life Sciences, Chemistry, Engineering, and Business. Also included is a helpful tutorial.

Currently high on the agendas of government oversight agencies and industry engineers is the evaluation and regulation of potentially toxic chemicals. The toxic, ecotoxic, carcinogenic, and other hazardous effects of these chemicals fall within REACH regulation in Europe and the Global Harmonization System in the U.S. “More than 90% of the approximately 30,000 chemical compounds currently on the market have never been appropriately tested for these unsafe effects,” stated Frank Lemke of KnowledgeMiner Software.

Given the constraints on time and cost, it is impossible to fulfill regulatory requirements without using alternative testing methods, such as computer-based QSAR models that can be easily developed by the industry itself, according to its changing needs. For example, a rodent carcinogenicity study of a single chemical compound has an estimated cost of over $1 million. Insights’ newest features allow users to develop and apply tailored, cost-optimized models from available data within minutes, resulting in massive cost and time savings compared with traditional animal testing. Mr. Lemke commented, “In addition, this approach can save the lives of millions of animals, and it opens the possibility of using powerful QSAR modeling right from the start of product development, in such areas as drug design, screening, and prioritization.” Employing user-supplied prevalence and cost matrices, or generating them automatically in software, Insights generates models that are optimal with respect to the given probability costs, and therefore allows developing both false positive- and false negative-optimized classifiers, for example.

Insights is a professional, yet convenient tool for building predictive models from data automatically. Taking observational data that describes a problem, system, or process, the software constructs a working mathematical model. Compatible with data stored in a variety of popular formats (e.g., Microsoft Excel), its AI-powered, self-organizing modeling algorithms allow users to easily extract new and useful knowledge to support decision-making. All models can be exported to Excel for further deployment.

Besides health and life sciences related problems, Insights can be applied to sales prediction, financial and resource planning, engineering problems, climate change, or mining collections of data from government agencies.

Language Support:
* English, German, and Spanish

System Requirements:
* OS X 10.7 or later
* Any Mac with 64-bit CPU
* Minimum screen resolution of 1280 x 768 pixel
* For Excel support, Excel versions 2011 or 2008
* 74.5 MB

Pricing and Availability:
Insights 2.0 is available as a free trial version exclusively from the KnowledgeMiner Software website. Insights 2.0 Ultimate, the high-end enterprise edition with cost-sensitive modeling, as well as an academic version, can be purchased by contacting KnowledgeMiner Software directly.

KnowledgeMiner Software
Insights 2.0
Download Free
App Icon

Located in Berlin, Germany, KnowledgeMiner Software was founded in 1993 by Frank Lemke. The company is active in research, development, consulting, and application of self-organizing modeling and knowledge discovery technologies. It developed and implemented a number of original technologies for validation of inductively built data mining models. KMS has been doing consulting in model development and prediction of toxicological and eco-toxicological hazards and risks of chemical compounds from experimental data for regulatory purposes within REACH and participated in three international research projects funded by the European Commission related to QSAR. Other fields of activity have been climate change related modeling and prediction problems, sales and demand predictions, macro- and micro-economic modeling problems like national economy and balance sheet prediction, energy consumption analysis and prediction, medical diagnosis, and wastewater reuse problems. Copyright (C) 2013 KnowledgeMiner Software. All Rights Reserved. Apple, the Apple logo, iPhone, and iPod are registered trademarks of Apple Inc. in the U.S. and/or other countries. Other trademarks and registered trademarks may be the property of their respective owners.