ANLY 530 Machine Learning – How Does Machine Learning Work? – Flashcards

an automated process that extracts pattern from data.

automatically learns a model of the relationship between a set of descriptive features and a target features based on a set of historical examples, or instances

to find the predictive model that generalize well

the set of assumptions that defines the model selection criteria of a machine learning algorithm

restriction bias and preference bias

contains the set of models that the algorithm will consider during the learning process.

guides the learning algorithm to prefer contain models over others. For example, in multivariable linear regression with gradient descent (Chapter 7), it implements the restriction bias of only considering descriptive features values and applies a preference bias over the order of the linear model it considers in terms of a gradient descent through a weight space.

occurs when the prediction model selected by the algorithm is too simplistic to represent the underlying relationship in the dataset between the descriptive features and the target features.

occurs when the prediction model selected by the algorithm is so complex that the model fits the dataset too closely and becomes sensitive to noise in the data.

One of the most commonly used process for predictive data analytics projects. It is non-proprietary, application and tool neutral etc. Key phases include:1) Business understanding including understanding of the business problem so we can design a analytic solution to address it; 2) data understanding: understand the data sources available and the kinds of data in the sources; 3) Data preparation: for example using ABT- Analytic Base Table ; 4) Modeling: build a range of prediction models from which the best model will be selected for deployment; 5) Evaluation: evaluate to be fit for the purpose; 6) Deployment: integrate the model into the process within the organization.