Predictive Analytics uses past data, current data, demographics information and other relevant data, to predict future events.
Prediction results are often given in probabilistic terms, for example the probability that a customer will respond to a new product offering, the expected donation amount to a charity, and so on.
Predictive Analytics encompasses a variety of techniques from statistics, machine learning, AI, data mining, optimization and others to analyze the relevant data in order to predict the future events.
Business applications of predictive analytics are:
The Analytic Engine mimics the way an experienced data scientist goes about in building predictive analytics model.
It is an expert system, a “predictive toolbox”.
If you are an experienced “data scientist” that like to tune the model parameters, you will be able to do so and add your domain expertise and data knowledge to the process.
If you are a beginner, or an “ordinary” business user, all you need to is 3 clicks to generate a model that holds years of “data science” experience in it.
From the uni-variate analysis to the feature selection and modeling process, we solve it in a fully automatic manner, resolving modeling and data issues on the fly, including non-linearity, outliers, missing values, numerical issues, convergence problems, statistical issues, decision issues and more.
With an easy to use process to load the data (rds or delimited files) and simple target setup and model selection, on the input side, and visualizing the modeling results and providing reports that yield a complete overview about the modeling process, on the output side, “Data science” is accessible more than ever.
Standalone Client Integrated (API)
Cloud (coming soon)
The DMWay Scoring Engine is responsible for deploying the Predictive Analytic results
It is the hhe heart of every predictive model.
Our scoring Engine serve to reduce days/weeks of integration process between the “data scientists” team and the operational environment that usually involved R&D, QA, IT and others. It avoids the frustration that data scientists often encountered seeing their model waiting on the shelf waiting for the IT people to execute the scoring process by shifting the power to the “data scientist” who, within a click away, can activate the scoring process. And not only this, but DM scoring engine can create the scoring code in several languages to fit the requirements of a diversity of organizations, including R, Java, Python, SQL and others.
Business models are dynamic and often age over time in light of new data and new conditions. The DMWay Re-calibration Engine is a satellite engine to the Analytic Engine. It gathers information and statistics about the performance and stability of the existing models to decides whether the model is still valid or becomes obsolete enough and thus need to be re-calibrated, either re-estimating the model parameters or rebuilding the model to reflect the new conditions. The re-calibration procedure is by no means simple and depends on the analytic model type. In linear regression models, for example, the re-calibration process starts with the existing model and updates the model incrementally based on the new data; in other models, such as logistic models, re-calibration often means building the model from scratch based on the new observations.
Contact us for further information…