Working with Profile Data
This chapter reviews the previous data sets in terms of what has been predicted: the price of a house in Iowa, daily train ridership at Clark and Lake, the species of a feces sample collected on a trail and a patient’s probability of having a stroke. Since the goal is to make daily predictions, the profile of within-day weather measurements should be somehow summarized at the day level in a manner that preserves the potential predictive information. Pharmaceutical companies use spectroscopy measurements to assess critical process parameters during the manufacturing of a biological drug. Models built on this process can be used with real-time data to recommend changes that can increase product yield. The chapter explores the performance of partial least squares (PLS), Cubist, radial basis function support vector machines, and feed-forward neural networks. PLS using derivative features appears to be the best combination.