Precision Alpha is scientific machine learning as a service that ingests customer time-series data and returns unbiased, model-free scientific measurements and ensemble forecasts to make optimal decisions.

Precision Alpha leverages machine learning and combinatorics to analytically move beyond statistics and determine a set of scientific measurements with closed-form expressions for any time series. No models are introduced, so there are no models to overfit.

Precision Alpha benefits all industries. Our solution is built on deductive combinatorial mathematics, the entire system independent of vertical, horizontal, or application. Precision Alpha web services expose behavior that is currently concealed in customer time-series, to benefit any industry.

Precision Alpha’s OnDemand web services are, by design and necessity, independent of vertical or horizontal. Their generality is their strength.

Financial Services:

Precision Alpha Exchange measures market emotions and a complete set of scientific measurements, determining bullishness or bearishness. This is accomplished by processing a historic time series of closing prices for every symbol on the NYSE and NASD every trading day four hours after the market close. Hedge funds with different trading horizons or asset class focus customize their input time-series using OnDemand.

Marketing:

Using Precision Alpha, small scale surveys can be utilized to gain insight into much larger demographics, often leveraging unique measurement tools to detect the emotions associated with survey data. Further, Precision Alpha can also leverage sales time series to optimize timing of advertising campaigns.

Inventory Management:

Precision Alpha processes historical sales demand for products at a store to calculate exactly how much product should be on the shelf to prevent lost sales opportunities. In a similar way, product manufacturing decisions are planned and monitored so that change in the inventory profile minimizes loss and better balances supply and demand.

Precision Alpha’s analytical machine learning is based on the science of counting, so that counted data is exact and the machine learning is combinatorial. Precision Alpha processes any time-series, to produce exact sensor measurements that are used to understand the non-equilibrium dynamics of the time-series.

Precision Alpha performs far better than standard forecasting methods (like a Monte Carlo simulation) because other methods assume all distributions are statistically normal, regardless of what the data indicates.

Precision Insight is designed to handle large, complex datasets. Precision Insight’s easy-to-use API allows submission of multiple time-series, even time-series using the same timestamps, so processing occurs together for easy comparison.

Precision Insight clients connect directly with our expert team, supplying a time series that can be processed and analyzed. Depending on the scope of the project, an initial analysis may be done as a proof of concept to gain visibility and insight into the time-series dynamics. Upon full subscription, clients are integrated into our system and gain access directly to Precision Insight’s S3 buckets.

Precision Insight integration is achieved through AWS’s security infrastructure (IAM), using S3 buckets and lambda functions to implement the services. No personally identifiable information or meta-data are communicated, and all processed files and output data are deleted permanently from the S3 buckets after one month.

Precision Insight’s mathematical innovations and technology stack were developed by co-founder and Chief Data Scientist, Mark Temple-Raston. Mark has a doctorate from Cambridge University in Applied Mathematics and Theoretical Physics, the same department as Stephen Hawking.

The Precision Insight team has over 30 years of experience in consulting and software development in aerospace, logistics, pharma, and financial services. The team also has over 20 years on Wall Street, including 10 years at Citigroup building enterprise compliance systems for Basel II and Dodd-Frank for data management.