Abraham Meidan, CEO
Organizations now have the ability to analyze their past performance and predict what lies in the future with the advent of big data and technologies such as machine learning. And to make most of such capabilities, New York-based WizSoft has unlocked several new ways to utilize and mine data. As the developer of WizWhy, the only program that reveals if-and-only-if rules explaining the values of the dependent variable, WizSoft simplifies data mining and predictive analytics for businesses.
WizWhy is an innovative data mining software that reveals the patterns in the data and offers the ability to explain how the values of one field in the data are affected by the values of other fields. Following the common statistical jargon, the field selected by a user to be explained is the dependent variable whereas the other fields are the independent variables. WizWhy’s capability to reveal if-and-only-if rules helps segment the rules that relate between the dependent and other variables. These rules explain the dependent variable and issue predictions for new cases. In an era when businesses seek advanced solutions to analyze large volumes of data, WizWhy presents easy-to-understand explanations for each prediction. “We expect there to be a demand for machine learning tools that can analyze data sets having more than 20,000 columns, and this demand will be mainly in regard to the analysis of genome data. We developed WizWhy, a data mining software tool, with these trends in mind, to support these features,” says Abraham Meidan, CEO of WizSoft.
The company’s WizWhy is perfectly aligned with organizations’ requirements for machine learning programs that present an easy-to-understand explanation for each prediction.
It can analyze data sets having up to 30,000 columns and an unlimited number of rows. Another factor that differentiates WizWhy is the fact that it issues accurate models even when the datasets suffer from many cases of missing data. One of the cases where this machine learning software delivered significant outcomes is during the company’s association with National Institutes of Health (NIH), a client that entrusted the task of analyzing human data to WizSoft. The WizSoft team then converted the data and built a table where each column referred to a protein-codegen and each row referred to another subject. The dependent variable was a certain sickness, and the content of each cell of the independent variables was the SNP (single-nucleotide polymorphism) that denotes a variation in a single nucleotide at a certain position in the genome. In this case, WizWhy revealed short if-and-only-if rules that explained the sickness as a result of the SNPs. WizSoft has had many such associations that speak volumes of the excellence it can deliver with WizWhy.
WizSoft does not compete with other companies and rather recommends that its algorithm can be used on top of those offered by other companies
Another key highlight of WizWhy is that it does not compete with other algorithms and rather recommends that its algorithm can be used on top of those offered by other companies. In Meidan’s opinion, contrary to other applications such as accounting or CRM, where it makes sense for an organization to use only one application, in the field of machine learning and data mining, it is appropriate to use several applications as each may reveal interesting patterns. WizSoft thus suggests businesses employ several algorithms and determine the final prediction by voting. This often helps to yield precise results, which is the ultimate goal for any organization while analyzing data. For the WizSoft team, it is all about providing the most accurate predictions. “We intend to continuously improve our algorithm,” concludes Meidan.