Enterprise-wide data mining software and services

Regardless of how expansive and well-coordinated data generation and databasing effort exists, the mere existence of massive and integrated data does not guarantee the ability to mine this data to address specific researcher needs. In many cases the data may not be suitable for certain types of exploratory or confirmatory data analysis and meta-analysis. There exist several data limitations (e.g., sample size, observation biases, confounding, dimensionality, etc.) that may make such analyses problematic. In addition, traditional visualization and analysis methods offered by data mining packages neither can perform such analyses (e.g., such tools typically overlook issues such as overfitting the analysis by manual data dredging, statistical unreliability of visually-derived conclusions and difficulty of conducting complex multivariate analyses without extensive expert analyst time investment). Most importantly such analyses are traditionally very time consuming and thus constitute a significant research bottleneck. At the core of our strategy for attacking to the problem is the creation of the Discovery Mining Engine (DME). This data analytics software platform is capable of enabling a wide range of hypothesis-driven and exploratory analyses using both clinical as well as high-throughput molecular data.

For questions and additional information, please contact Alexander Statnikov.