Project for Coursera Developing Data Products class

Problem: Classify Poorly-Segmented (PS) vs. Well-Segmented (WS) objects in cell screening images.

Link to online shiny application

• Image processing segmentation problem
• Shiny interactive web apps
• Caret package machine learning models

• Preprocessing options and Caret model selection at left
• 3 dynamic panels and static help panel on right

ConfusionMatrix panel shows confusion matrix, accuracy, sensitivity, specificity and other statistics.

Let's look at % Accuracy.