SAS Predictive Modeling - (3 in 1) with 3 live Case Studies

Discover what business analytics is and how it drives insights from data. Learn its distinctions from business intelligence and explore applications across finance, marketing, supply chain, and human resources.

Explore the three main analytics categories—descriptive, predictive, and prescriptive—and how each uses data to describe, forecast, and optimize business outcomes.

This lecture explains linear regression, including simple and multiple forms, independent variables, the line of best fit, and the standard equation; it covers adjusted r-squared, f-statistic, and p-values.

Apply practical linear regression to predict customer satisfaction using significant variables such as product quality, price flexibility, and packaging, validate predictions on training and validation data.

Explore the assumptions of logistic regression, where the dependent variable is binary and independent variables can be continuous or categorical, with probabilities between 0 and 1; assess goodness of fit.

Explore odds and odds ratio within logistic regression, linking a binary dependent variable to predictors, and explain how changes in x affect odds through the logit and sigmoid forms.

Set the cut point probability level in logistic regression to convert predicted probabilities into binary outcomes (0/1), using a credit risk example to distinguish events from non-events.

Discover how confusion matrices evaluate classification models, calculating accuracy and exploring true/false positives and negatives, along with sensitivity, specificity, and area under the ROC curve.

Explore logistic regression for a binary loan repayment outcome, interpreting odds and odds ratios, evaluating models with confusion metrics and roc curves while applying a credit dataset and case study.

Explore logistic regression in credit risk analytics, three live case studies, selecting significant variables via stepwise methods, estimating maximum likelihood, and predicting loan default probabilities with SAS Proc Logistic.

Apply practical logistic regression to credit risk data, convert estimated probabilities to binary status with a 0.5 cutoff, and identify significant variables for model building.

Explore logistic regression in SAS applied to credit risk analytics, selecting significant predictors, computing predicted probabilities, and evaluating model performance with confusion metrics, thresholding at 0.5.

Explore the components of time series analysis, including secular trend, cyclical, seasonal, and random fluctuations, and learn how these elements shape forecasting and business planning.

Explore smoothing techniques in time series data to reduce random variations and reveal underlying trends. Learn moving average and exponential smoothing, their weights, and applications to stock prices and seasonality.

Explore stationarity and non-stationarity, the random walk concept, and how differencing and transformations convert data to stationary for forecasting.

Explore arima modelling for time series analysis, focusing on making data stationary through transformations and differencing, and combining autoregressive and moving average components to forecast future values.

Explore time series concepts, including trend, seasonality, and stationarity, and apply arima and moving average techniques to forecast airline passengers from 1949–1960 for the next year.

Practice time series modeling with an airline dataset containing date and passengers, test stationarity, and forecast the next 12 months using SAS procedures.

Apply time series methods to predict passenger travel for the next two months using differencing, stationarity checks, and hypothesis testing in SAS predictive modeling.

In this hands-on time series session, students model and forecast air passenger counts from 1949–1960 using data prep, differencing, and information-criterion based model selection with confidence interval forecasts.