3 Experimental Results
3.1 In the first round, preliminary results are small in quantity and tainted by an evolving experimental protocol from one subject to the next. Processes were refined, training was standardized, and we saw the variance between human classifiers decline. Qualitatively, we show enough to contend that the project will be productive going forward.
3.2 Use Cases
3.2.1 Synthetic Data generators have been built, and simple problems of overlapping Gaussian distributions have been modeled.
3.2.2 GvHD report
3.2.3 SIV report
3.3 Algorithms
3.3.1 Magnetic Gating Report
3.3.2 Probability Bin Clustering Analysis Report
3.3.3 WEKA ANN Report
3.3.4 Support Vector Machine Algorithm Preliminary Report
3.3.5 Automated Classifier Report
3.4 Metrics
3.5 Infrastructure
3.6 Phase II Application Documents
3.7 Lifecycle Architecture Milestone