Predictive Analytics & Modeling: R | Minitab | SPSS | SAS

Explore a propensity model to predict card purchases by importing a dataset, examining variables like gender, country, income, and scores, and evaluating model performance with event rate.

Explore IV calculation in R by checking missing values and unique levels, computing information value with weight of evidence, and using backward elimination to refine a logistic regression model.

Explore categorical variables with ggplot2 by creating bar plots that compare card offer across gender and country region, encode categories, and interpret event rates around 15%.

Split data into training and test sets with caTools, then explore binning via decision trees and weight of evidence binning, optionally scale numeric variables, and fit logistic regression.

Apply backward elimination with a GLM to identify key variables: country, income, holding balance, and credit score, and build an optimal model assessed by AIC and ROC AUC.

Explore building a lift chart or gain chart for the training set to evaluate predictive models, using deciles and quantiles to assess score distributions and performance.

Explore decile-based model evaluation with lift and gain charts, calculate event and cumulative percentages for goods and bads, and assess training versus test set performance to refine targeting.

Assess model performance on the test set by computing test scores, analyzing lift in the fourth decile and case, and evaluating out-of-time stability.

Save your logistic regression model in R and load it for scoring new data. Compare training and test performance, decile lift, and scores to assess model quality.

Fit a decision tree model to compare its performance with logistic regression, noting that scaling is not required and missing values are handled automatically in training and test sets.

Fit and interpret a decision tree classifier using splits like preferred customer score < 0.5546 and estimated income to predict purchases, and visualize the model with plotting options.

Learn to generate predictions with a decision tree, interpret class probabilities, evaluate model performance with a confusion matrix and accuracy, and explore pruning and forests for better results.

Explore the overview and history of R, from the S language roots to an open-source statistical programming language with CRAN and Bioconductor packages and active community.

Explore explicit coercion in R with as dot star functions and NA results, then create, convert by dimension change, inspect, and bind matrices using matrix, dim, and cbind/rbind.

Explore data types and basic operations in R, including numeric, character, integer, complex, and logical objects, vectors, lists, missing values, data frames, factors, and attributes, with assignment and print functions.

Explore evaluation and printing in R, including auto vs explicit print, console, and vector creation with colon notation and c() function across numeric, logical, character, integer, complex types, and coercion.

Handle missing values in data with R by distinguishing na from nan, testing with is.na and is.nan, and working with data frames, vectors, and matrices.

Explore data types in R such as numeric, logical, character, integer, and complex; examine vectors, lists, factors, missing values, and data frames, and learn how to name objects.

Master subsetting in R for lists and matrices using single, double brackets and dollar, explore partial matching and vectorized operations, and use drop to preserve matrix form when extracting values.

Explore subsetting nested lists and vectors in R, including dollar versus double-bracket indexing, partial matching, removing any values with complete.cases, and perform vectorized and matrix operations.

Set the working directory in R and read a sample txt file into a table; R infers types, places data in columns a to e, and skips hash lines.

Improve data handling in R by using read.table to load large tables, estimate memory, set colClasses to speed up reads, and load only top rows or in parts.

Assess system capacity for large data in R by estimating memory needs, confirming RAM and 64-bit OS, and using dput, dump, and dget for version-control friendly data.

Learn to read and write data in R using dump and read lines, create and manage connections to local files and URLs, and handle gz compressed data.

Explore control structures in programming, including if else, for loops, while loops, repeat, break, skip, next, and return, with syntax and flow-control examples.

Define functions in R as first-class objects, show their syntax and how to pass, nest, and return values. Explain formal arguments, named versus positional matching, defaults, and partial matching.

Explore defining functions in R, including default and null arguments, lazy evaluation, and error handling when arguments are missing; learn how ellipsis and triple-dot syntax enable flexible argument passing.

Explore how R binds values to symbols through lexical scoping, the search list, global environment, and namespaces, and why package order affects lookups.

Explore how R environments store symbol-value pairs and resolve free variables. Understand closures and lexical scoping through nested functions in practice.

Explore how function environments drive variable lookup with lexical and dynamic scoping in R, illustrated by y, f, and g. See how lexical scoping enables efficient optimization of likelihoods.

Discover looping in R with lapply, sapply, apply, tapply, and mapply, using split and anonymous functions to drive results and understand when lapply yields a list versus sapply's simplification.

Explore how to apply functions across matrix margins with apply, tapply, and split in R, computing means, sums, quantiles, and group ranges for predictive analytics.

simulate a linear model y equals 0.5 plus two x plus e with normal, binomial, and Poisson inputs, and use set.seed and sampling to ensure reproducibility.

Learn to create plots in R using base graphics, lattice, and grid, including hist and box plots, and to choose devices (X11, pdf, png) and customize with par parameters.

Explore ggplot2 and the grammar of graphics to map data to aesthetics like color, shape, and size, and compare ggplot2 with base and lattice plotting systems in R.

Explore how to create and customize plots with qplot in ggplot2, including point and smooth lines, histograms, facets, and density curves for grouped data.

Explore ggplot2 as an implementation of the Grammar of Graphics, building layered plots from a data frame with aesthetic mappings, geoms, facets, scales, and coordinates.

Master ggplot2 basics by annotating plots with x lab, y lab, labs, and gg title, and by customizing themes, colors, and point aesthetics while illustrating outliers and smoothing with lm.

Master regex basics by using meta characters to mark line starts and ends, character classes, and alternation with | to extract patterns from text like social media feeds.

Master regex patterns match terms like flood or earthquake, extract sentences starting with good or bad, apply alternation, grouping, star, and optional metacharacters, with slash escape to treat dots literally.

Explore how regular expressions, with metacharacters like star, plus, and brackets, enable repetition and data extraction in text analytics, useful beyond R.

Explore object oriented programming in R, focusing on S3 and S4 classes, the methods package, and how set class, set method, and set generic enable method dispatch.

Explore how generic functions dispatch to S3 and S4 methods based on an object's class, including default and trace methods, and how get method and get S3 method retrieve code.

Practice part 2 debugging by writing test cases, comparing expected and actual results, reproducing the problem, and using traceback, debug, browser, trace, and recover to step through code.

Learn to use Minitab for descriptive and inferential statistics to study data and support sound business decisions. Understand population, sample, frame, gap, and the difference between non-probability and probability samples.

Explore data types, including attribute (categorical or count) and measurement (continuous), with nominal, ordinal, interval, and ratio scales, then learn mean, median, mode, range, and interquartile range.

Explore descriptive statistics in Minitab, computing mean, standard deviation, variance, range, quartiles, and interquartile range from neck, BMI, and body fat data, with grouping by gender for comparison.

Sort data in minitab to organize a randomized dataset of 100 rosewood high students by gender, ethnicity, and body mass index, storing results in new columns.

Explore how to create a histogram in Minitab with 36-month data set of working days. Learn to add data labels, adjust titles, colors, and display class intervals with cut points.

Learn to create and customize pie charts in Minitab, displaying ethnicity and subject distributions with slice labels and colors, and saving graphs for external use.

Create bar charts in Minitab to display counts of categorical data, like ethnicity and subjects, add data labels, edit titles, customize colors, and chart data from tables.

Learn how to create line graphs in Minitab by plotting graduation year against the percentage of students admitted to tier one universities, revealing time-based trends.

Compute the mean and standard deviation of a discrete random variable using a binomial distribution in Minitab, and determine the expected value for a group of 14.

Explore binomial distribution with minitab by computing event probabilities for a 15-birth dataset, including at most ten and at least twelve boys, using a 0.5 success probability.

Explore normal distribution probabilities with minitab, using a mean of 95.7 and sd of 4.9 to compute areas below 89.6, above 102, and 5% tails for fasting blood sugar levels.

Learn how to check normality in Minitab using three methods on lab testing time data. Validate normal distribution with Anderson-Darling test, probability plots, and normality test, interpreting p-values around 0.119.

Explore transforming non-normal data into normal using the Box-Cox method in Minitab, illustrated with baking time samples; test normality with Anderson-Darling and confirm transformed data follows a normal distribution (p>0.05).

Minitab demonstrates generating a suitable sample from a class of 100 students' gender, ethnicity, and BMI data and determining the ideal sample size for analyses.

Learn how to determine sample size for estimation of proportions and means using minitab, guided by examples with confidence levels, margin of error, historical data, and planning values.

Learn parameter estimation with Minitab by calculating proportions, means, and standard deviations, and constructing 95% confidence intervals using exact, chi-square, and bonnet methods.

Explore power analysis with Minitab for proportions, means, and standard deviation, using real IVF, heart-rate, and processing-time examples to determine sample sizes and test power.

Apply measurement system analysis in a Six Sigma project using Minitab, ensuring valid data before decisions. Explore gage R&R for continuous data, covering accuracy, repeatability, reproducibility, linearity, stability, and thresholds.

Conduct a Gage R&R study in Minitab using ANOVA to analyze parts and operators, read R and x bar charts, and evaluate variation and distinct categories for acceptance with caution.

Assess a measurement system for discrete data with attribute agreement analysis in Minitab, using multiple appraisers and a standard to evaluate within-appraiser, between-appraiser, and team accuracy.

Analyze process capability with cp, cpk, pp, and ppk for continuous data in Six Sigma, using Minitab; the example indicates the process is not capable in the short term.

Apply the paired t test in minitab to compare dependent means, burger patty frying times with and without the additive, testing whether the difference exceeds seven minutes at 95% confidence.

Explore how the Pearson correlation coefficient measures the linear relationship between calories consumed and weight gained, interpret r from -1 to 1, and assess statistical significance with p-values in Minitab.

Link regression to correlation by deriving a simple linear model y = b0 + b1 x + e from historical data to predict future y and measure fit with r-squared.

Learn how control charts drive Six Sigma improvements by distinguishing common and assignable variation, monitoring attribute and variable data, and catching out-of-control conditions before defects.

Apply one-way anova and basic regression concepts, explore scatter plots and regression analysis in minitab, and import data to perform descriptive statistics and graphical summaries.

Explore descriptive statistics, means, standard deviations, t tests, and skewness and kurtosis using Minitab, with practical data from mutual fund returns.

Explore descriptive statistics in Minitab to assess fund returns, including mean, standard deviation, variance, skewness, and kurtosis. Interpret how confidence level and graphical summaries inform risk comparisons and investment decisions.

Observe how standard deviation measures risk and volatility across funds, guiding investment choices to match a listener's risk appetite, with data organized using Minitab and Excel.

The lecture analyzes NAV price statistics, including mean, standard deviation, and range, noting higher volatility for ICICI Prudential Tech Fund, Banking and Financial Services Fund, and HDFC Equity Fund.

Assess fund volatility by examining standard deviation and range; identify X equity and HD cap as having the lowest volatility, while IBF and Itec show higher volatility.

Explore descriptive statistics in Minitab across finance, medical, and energy data, using mean, standard deviation, range, and skewness to interpret risk and volatility.

Apply descriptive statistics to customer complaints and resting heart rate data to highlight mean, median, standard deviation, and skewness, and interpret three shifts and activity effects.

Compare before and after resting heart rate using descriptive statistics, noting minimal mean and median differences, and highlight data quality, interpretation, and loan applicant skewness for predictive modeling.

Analyze loan applicant data by examining income distribution, education level, age, savings, and debt, noting high income variance, negative skewness, and typical credit card ownership.

Analyze loan applicant data through predictive analytics, noting income variability, high savings dependent on spending, low debt, and credit card usage up to six, with R, Minitab, SPSS, and SAS.

Explore the features of the t test in predictive modeling, including single-sample and two-sample t tests, p-values, and interpretation using resting heart rate data in Minitab.

Use a paired t test in Minitab to assess if debt depends on income for a loan applicant, interpreting t and p values from the income and debt data.

Apply paired t tests and two-sample tests to determine if savings affect debt, interpreting t and p values in predictive modeling with Minitab, SAS, SPSS, or R.

Explore one-way ANOVA to determine if mutual fund return means differ, using minitab to compute p-values, r-squared, and confidence intervals for hypothesis testing.

Assess pairwise comparisons in anova by evaluating p values, r-squared, and confidence intervals; conclude that not all means are equal when p < 0.05, rejecting the null hypothesis.

explains how to compare observed and expected frequencies using a chi-square (g square) test, including degrees of freedom, null and alternative hypotheses, and p-values, with a practical umbrella handles example.

The instructor explains g square chi-square tests to compare observed versus expected frequencies for smoking preferences and pulse rates before and after running, with 95% confidence and p-values.

Analyze differences between growth and dividend plans in mutual funds by comparing nav and repurchase prices, and test observed versus expected prices using chi-square tests.

Illustrates using chi-square goodness-of-fit in minitab to compare nav price and repurchase price, test null vs alternative hypotheses, and interpret p-values and critical values to decide on rejection.

Explore basic correlation techniques to understand positive, negative, and zero relationships, interpret correlation coefficients between -1 and 1, and apply these concepts with sample data in Minitab.

Explore how to compute Karl Pearson's and Spearman's rho in Minitab, build a store matrix from data, and understand why the unitary matrix omits diagonal correlations in a 4x4 matrix.

Continue on implementation using Minitab demonstrates arranging variables, exploring correlations, and interpreting values like 0.853, 0.778, and 0.015, with color-coded results in predictive analytics.

Interpret correlation values within a 5x5 matrix, highlighting diagonal 100% correlations and positive or negative links. Learn how these patterns inform diversification and predictive modeling with Minitab.

Calculate Pearson correlations with Minitab to distinguish positive, negative, and zero relationships in mutual fund returns. Note self-correlation of 100% and the use of r and p-values in interpreting associations.

Explore how correlation values reveal diversification benefits across sectoral funds, with AI tech and IBF showing strong diversification; learn credible predictive interpretations for portfolio decisions.

Explore how resting heart rate varies before and after rest using Pearson correlation, revealing a strong positive relationship (r = 0.716) between the two measures.

Interpret heartbeat variations before and after rest and practice deciding when prediction is meaningful, then perform correlation analysis on income, savings, and debt in a demographics dataset.

Analyze correlations among income, savings, and debt using a tabulated matrix, noting a positive correlation between income and savings and negative correlations between income and debt and savings and debt.

Analyze observed correlations among income, savings, and debt, noting a positive link between income and savings and negative links with debt, organized in tabulations for demographics and living standards.

Explore scatterplots with regression to analyze relationships, compute correlation values such as positive 0.21 between income and savings and negative correlations with debt, using multiple graphs for panels.

Explore scatterplots with regression to reveal strong positive correlations across heartbeat data and HDFC equity, and interpret correlation values such as 71.6% in predictive analytics and modeling.

Explore regression modeling from simple linear regression y = mx + c to interpreting r squared, t values, and p values, and predict outcomes using Minitab.

Identify the independent variable as weight (a continuous predictor) and the dependent variable as heartbeat after run, to fit a regression model in Minitab, interpreting the regression equation and r-squared.

Tabulating these values highlights relevant variables, noting the y intercept t value is 8.93 with a zero p value, and weight is insignificant for the after run heart pulse.

Analyze how a smoker's heartbeat depends on weight using a regression equation, interpret t and p values, and predict pulse from weight with Minitab and Excel.

Explore how smoker weight affects post-run heartbeat, with higher weight linked to lower heart rate and a negative weight coefficient. Assess regression significance and fit for before-run and after-run heartbeats.

Explore how regression outputs are written as y = mx + c and interpreted alongside r square and p values, with weight shown as insignificant and descriptive statistics for height.

Compute corresponding y values from given weights and heights to assess before-run smoker pulse, then use Minitab to create scatter plots with regression lines and interpret weight- and height-related trends.

Identify dependent and independent variables in a regression model, showing how energy consumption depends on machine energy settings, and fit a simple linear regression in Minitab to derive the equation.

Explore descriptive statistics for machine setting as the independent variable and its impact on energy consumption, including mean, min, max, range, standard deviation, and a regression relation.

Explore scatterplot analysis and regression modeling to assess energy consumption relationships and copper expansion with temperature, using minitab to interpret r-squared, p-values, and model fit.

Analyze a simple regression model that explains about 69% of the variance, using Kelvin as predictor with y = 0.021060 Kelvin + 7.449, and significant t and p values.

Explore p-value and t-value in the context of simple linear regression modeling copper expansion as a function of temperature in kelvin, using Excel to compute the regression equation.

Explore how temperature changes drive copper expansion through regression, with kelvin as predictor, expansion as dependent variable, and a scatter plot showing a 0.83 correlation and 68.95% r-squared.

this lecture uses a finance example to test whether Reliance and Infosys stock returns depend on BSE Sensex returns, using Minitab to fit regression and report r-squared and p-values.

Analyze example 5 interpretations, noting R square 50.16% and Sensex returns predicting Reliance returns with significant t and p values, and apply the regression equation in Excel for unit-term predictions.

Apply the regression equation to generate predicted Infosys returns from BSE Sensex changes, using Excel, and interpret results with R-squared, t and p values, plus a regression scatter plot.

Explore linear and simple regression, interpret R-squared, p-values, and t-values; create regression equations and scatter plots for Reliance and Infosys against Sensex.

Explore how density and temperature influence the stiffness of a plastic board through multiple regression, using Minitab to model predictors and interpret the t, p values and r-squared.

Explore continuing multiple regression with an example predicting stiffness of a plastic board using density and temperature; interpret regression coefficients, r square, and p values, noting density as significant.

Explore multiple regression with density and temperature predicting stiffness in a plastic board, using Minitab to identify the dependent variable, predictors, and model statistics.

Build a basic regression model predicting stiffness from density and temperature, compare models with and without temperature, and interpret confidence intervals, t and p values, plus a scatter plot.

Identify the dependent variable y as total heat flux and model it with predictors insulation east, north, south, and time of day using regression, checking multicollinearity and noting r square.

Explore regression outputs and observations, interpret p-values for predictors (insolation, east, north, south) and time of day, and interpret heat flux implications with an r-squared of 89.88%.

Time of day remains insignificant in the current regression model; the lecture covers two regression equations (with and without time of day) and Excel-based descriptive statistics to compare predictions.

Analyze how insolation and time of day shape total heat flux using a regression model and predict values from the regression equation, with and without time of day.

Generate scatter plots of heat flux against insolation, east, south, north, and time of day, interpret regression outputs, note correlations and multicollinearity, and apply to cotton wrinkle resistance.

Analyze how formaldehyde concentration, catalyst ratio, temperature, and time affect cotton's durable press wrinkle resistance through a four-variable regression in Minitab, yielding a 72% R-squared and an interpretable regression equation.

Explain regression outputs by interpreting r-squared 72.98%, identifying significant predictors like concentration and temperature, noting insignificant constants and time, and evaluating p-values, t-values, and f-values for best-fit models.

Learn to build a regression model predicting wrinkle resistance rating from concentration, ratio, temperature, and time; assess intercept and variable significance and explore 90% and 75% confidence intervals with Excel.

Display descriptive statistics by showing min and max values for concentration, temperature, and time, and show how inputs drive predicted values in predictive modeling.

Explore scatterplots and regression analyses for ferrite, aluminate, silicate, and trisilicate, comparing simple and multiple regression amid multicollinearity, with r-squared and p-value interpretations.

Use a regression-based approach to calculate density from known temperature and stiffness by adapting the regression equation, enabling density estimates for specified stiffness and temperature values.

Explore how to compute a specific independent variable in a multiple regression model to predict density across varying temperatures and stiffness, showcasing predictive modeling and logistic regression next session.

Explore logistic regression with dichotomous and categorical variables, modeling how smoking affects after running heart pulse across gender, using height and weight as predictors.

Extend regression analysis across gender by deriving separate equations for females and males, using height and weight to predict after running heart pulse, with Minitab demonstration on a 90-respondent dataset.

Generate and interpret regression equations for heart pulse using height and weight as continuous predictors, with gender and smoking as predictors, evaluating model fit with p-values, t-values, and r-squared.

Explore tabulated values and regression outputs to determine variable significance and predictability, highlighting ambiguous outputs, weak r square, and lack of strong correlation in scatter plots.

Apply regression modeling to predict sales from client count and years, using group-based dummy variables and separate equations to compare three company segments in Minitab.

Analyze regression results across three groups, interpreting r-squared at 81.69%, the t-values and p-values, and the role of clients versus years in business, while noting multicollinearity.

Interpret regression results with an r square of 81.69%, assess variable significance via t tests, and show how sales rise with client count and years in business across three groups.

Interprets regression-based sales predictions by analyzing how client counts and years affect revenue across three groups, using Excel equations to estimate and compare projected sales.

Examine how a regression equation uses clients and years in business to predict group sales across three groups, illustrating how changing inputs alters the predicted values.

Interpret how to plot sales against clients and years across groups, analyze scatter plots with regression, and interpret positive correlations despite limited data points.

Learn to implement scatter plots and generate predicted values, handling categorical predictors and regression outputs, including R-squared, t-values, and p-values, with examples using temperature, strength, and manufacturers.

Learn how to model plastic case strength as a function of temperature with manufacturers as a categorical predictor, derive regression equations for each manufacturer, and interpret r-squared around 0.59.

Analyze separate regression equations for manufacturers a and b, with significant constant and temperature, r square indicating best fit, and exploring predicted strength values.

Generate predicted strength values for manufacturers A and B in excel using the equation 7349 minus 11.47 temperature, and show how temperature between 183 and 208 shifts strength.

Analyze a scatter plot of plastic strength versus temperature for manufacturers A and B; manufacturer B declines sharply with heat, while manufacturer A preserves strength better.

Explore how logistic regression analyzes cereal purchase decisions using income, whether viewers have children, and ad exposure, yielding four equations to predict buying behavior.

Develop and interpret regression equations linking income and ad exposure to children viewed across four scenarios, using Excel to format, compute, and compare predicted outcomes.

Explore how to derive predicted values for individual customers across four regression equations conditioned on income, ad exposure, and parental status, and interpret R square, t values, and p values.

Evaluate regression results by examining income as an independent variable, interpreting p-values, t-values, and R-squared to assess significance and predictive ambiguity, with scatter plots illustrating outcomes.

Apply logistic regression to determine how age, education, debt, and savings influence income and guide credit card granting decisions for users and non-users.

Explore a regression-based credit card grant decision, using tabulated values and outputs like r-squared and p-values to evaluate age, education, debt, and savings as predictors.

Explore how regression interpretations vary across credit card scenarios, with changes in education, age, and savings shaping income, while assessing intercept shifts and predictive value using Excel.

Explains building a predictive model for credit card approvals across four situations (nn, ny, yy, yn) using age, education, savings, and debt, and highlights debt's insignificance.

Explore how to interpret scatterplots in predictive analytics, comparing predicted values with debt factor scenarios to understand income level effects and positive or negative correlations.

Develop predictive modeling in Excel with the Data Analysis Toolpak, covering ANOVA, t-test, F-test, regression, correlation, and descriptive statistics.

Activate the Excel data analysis toolpak and run descriptive statistics for height and weight. Set input ranges and outputs, apply a 95% confidence level, and format results to two decimals.

Learn how to implement single-factor ANOVA in Microsoft Excel using the Data Analysis Toolpak, including setting alpha to 0.05 and interpreting p-values and F statistics for comparing multiple groups.

Learn to implement correlation in predictive modeling with Excel, using data analysis tools and the correlation formula to compute correlation coefficients between variables.

Implement linear regression in Excel using the data analysis toolpak, selecting y and x ranges, interpreting R-squared and ANOVA, and reviewing line fit plots and residuals for prediction tasks.

Apply descriptive statistics in SPSS to identify the highest and lowest crime incidences across states, using larceny, murder, and robbery examples like Arizona, Mississippi, Nevada, and New York.

Analyze descriptive statistics in SPSS using the gasoline dataset, importing data, selecting length and faults, and interpreting variance, standard deviation, skewness, and kurtosis to assess quality and normality.

Import text and csv datasets in SPSS, then apply descriptive statistics, correlations, and regression modeling for predictive insights.

Examine descriptive statistics for datasets, computing mean, standard deviation, variance, kurtosis, range, minimum and maximum, and interpret implications for volatility and returns in predictive modeling.

Explore descriptive statistics and graphing in SPSS using stock returns as examples, including data view, variable types, scatter plots, and setup for regression modeling in predictive analytics.

Explore descriptive statistics in SPSS, including mean, standard deviation, skewness, kurtosis, and range, and interpret the output from sample data sets.

Explore the basics of predictive modeling with SAS Enterprise Miner, create a new project, configure data sources, and navigate menus and nodes to build early models.

Select the rest data table, review its properties and metadata, and set the target variable for binary modeling in SAS Enterprise Miner, then create a process flow diagram.

Create an input data node from a data source, apply data partitioning and filtering for outliers, and connect nodes to build a predictive analytics workflow in SAS Enterprise Miner.

Configure metadata for time series analysis by defining fields such as month, product, state, and sales as target variable, creating a transaction data source, and applying seasonal decomposition.

Extend predictive modeling with SAS Enterprise Miner by adding more data sources, building exploration diagrams, and using stat-explore to assess chi-square and Cramer's v for variable importance.

Explore a three-variable data set—month-to-month saving balance changes, interest rate differential, and ads expenses—using Stat-xplore, Multiplot, and Graph Explorer to assess Pearson and Spearman correlations and generate interactive plots.

Explore visualizing a binary response with scatter plots, examine age and income as variables, and perform variable clustering and regression steps in predictive analytics.

Learn to run and interpret a cluster node, including standardization, Ward's clustering, segment plots, and associated statistics, plus a glimpse into variable selection and data partition steps.

Master variable selection and transformation in SAS Enterprise Miner for predictive modeling across continuous and binary targets with numeric or nominal inputs, including variable clustering and before/after transformations.

Demonstrates a SAS Enterprise Miner workflow from input and data partition to variable selection and regression, identifying a continuous interval target and selecting high-impact variables using r square criterion.

Explore variable selection in SAS Enterprise Miner, identifying five key inputs that drive the regression model and reveal their impact on the target variable.

Explore how to select input variables for a binary target using chi-square or r-squared criteria, with a bank email response example, data partitioning, and regression modeling.

Examine predictive analytics and modeling workflows, tracking selected versus rejected inputs, 12 of 282 variables, and assessing model performance with chi-square tests and variable importance visuals for binary targets.

Examine the variable frequency table across 19 clusters, including selector variables, cluster plots, and a dendrogram, then see regression modeling with stepwise selection and model comparison for credit line.

compare two regression models via variable clustering and a decision tree node, updating inputs from both paths to assess training, validation, and test performance with mean and max predicted.

Build a predictive analytics workflow by connecting a decision tree to a regression model, using variable clustering and selection to form leaf segments, and comparing models with output statistics.

Explore variable selection and fit statistics in predictive modeling, comparing regression and decision tree approaches, examining training and validation results and model comparisons.

Explore transformation of variables using the transform variables node to convert inputs into regression-ready features, and compare transforming before versus after variable selection in a binary target workflow.

Explore regression modeling with score rankings overlay, variable selection before and after transformation, and a final model linking inputs to outputs via lift, r2, and training validation metrics.

Update variable transformations with optimal binning and merged inputs, run a regression model to predict credit score, and review SAS code and training and validation results.

Compare and combine decision trees, regression, and neural networks to evaluate models against binary and ordinal targets, using a flow diagram, data partitions, and model comparison analysis.

Analyze how a binary output variable is modeled using decision trees, neural networks, and regression methods, and interpret predictions, node rules, variable importance, and training versus validation performance.

Review the regression node outputs for a normal target with loss frequency; examine parameter estimates, standard errors, and fit statistics, then explore the decision tree results and SAS code execution.

Analyze a subseries plot with the attrition target variable's variable importance and fit statistics, and compare decision tree and gradient boosting using ROC curves and SAS code.

Create an ensemble diagram by combining logistic regression, decision tree, and neural network models with an ensemble node, using the attrition target and data partition and impute steps.

Explore decision tree modeling in SAS Enterprise Miner to predict a binary response and a continuous loss frequency, using data partition node, profit matrix, and SAS code node.

Run and update a decision tree model on partitioned data, interpret input and target variables, review fit statistics and tree structure, and execute SAS code.

Apply the score node to a prospect dataset, predicting the probability of response with a decision tree, data partitioning, and model comparison in the D score diagram.

Analyze a regression tree model's output, depth, and leaf statistics. Compare training, validation, and test results within SAS Enterprise Miner.

Interactively build and modify a decision tree workflow from a root node, using an input node, partition node, three decision trees, and a model comparison to evaluate results.

Change the data split and rebuild the model to produce an interactive decision tree, showing target and input variables, leaf nodes, and fit statistics across train, validation, and test sets.

Explore creating and refining interactive decision trees: split nodes via right-click, apply nominal and ordinal rules, and train branches from scratch in the interactive window.

Build a decision tree from scratch in Enterprise Miner by iteratively splitting nodes using entropy and log values, handling missing data, and expanding to the maximum tree; assess performance.

Explore neural network models in SAS Enterprise Miner to predict response and risk for auto insurance, using a two-layer network with hidden units and probabilistic outputs.

Learn how a neural network model runs and is evaluated, from binary inputs and a binary target to training iterations, average squared error, misclassification rate, and SAS-based model comparison.

Explore how a neural network derives optimal weights across iterations by viewing weights history and final weights, noting the average squared error on the validation set peaks around iteration 49.

Explore how a neural network outputs final weights and performance metrics, including iteration weights, error plots, and SAS code generated by enterprise miner, with emphasis on binary target prediction.

Score and compare neural network models with SAS, performing model comparison and scoring, and examining fit statistics, lift, and cumulative performance across training, validation, and test data.

Explore the neural network iteration plot to inspect training metrics such as average and root mean square errors, misclassification rates, and weight histories, with SAS and RSP scoring insights.

Compare neural network results in SAS code by examining lift, gain, percentage response, and errors, and interpret final weights and model history across iterations.

Explore how changing neural network models affects cumulative lift and related performance metrics, comparing eight networks through model comparison plots, SAS code, and scoring across training, validation, and test sets.

Compare DM neural, auto neural, and DM mine regression nodes by building a diagram, configuring data, targets, weights, and partitions, then run and compare models.

Explore the auto neural node results, including weights, iterations, and training settings, in SAS Enterprise Miner. Review the generated SAS code, scoring, lift, gain, and model diagnostics to assess performance.

Compare DM neural, deep neural regression, and auto neural; auto neural delivers the lowest average squared error, lower misclassification rate, and highest cumulative captured response and profit.

Create a binary target variable with prior probabilities, partition data, and build multiple neural networks (multilayer perceptron, radial width variants, auto neural) to compare performance.

Examine how switching to the average squared error as the error function alters iteration plots, misclassification rates, weights history, and SAS-backed neural network results.

Explore how the score rating overlay analyzes iteration plots, error metrics, and weight updates from the auto neural node to reveal cumulative lift, gain percentage, and SAS code generation.

Run the d mine regression node and review outputs such as training proportions, predicted variables, fit statistics, and ROC comparison, noting the de mine regression model performs best.

Explore regression models with binary and continuous targets using SAS Enterprise Miner, including regression node properties like link function, selection model, and criteria, applied to mail campaign response prediction.

In regression analysis with SAS enterprise miner, learn to read table effect plots, interpret intercepts and effects, and assess lift, gain, and cumulative response for binary and ordinal targets.

Explore building and evaluating a regression model to predict ordinal and nominal targets using Enterprise Miner, covering data partitioning, model diagnostics, and SAS code generation.

Create a flow diagram to build a logistic regression with backward selection in SAS Enterprise Miner, using RSP as target, partitioning, training/validation/test sets, and reporting odds ratios and SAS code.

Explain how predictive modeling uses historical and current data, applies techniques, and generates futuristic data. Identify trends and patterns, and recognize meaningful information as data used for prediction.

Explore predictive modeling, a machine learning approach that uses statistical and mathematical techniques to transform data into models that produce accurate future estimates, patterns, and trends.

Learn to build a predictive model by assembling customer attributes into a dataset, plotting two variables like age and items purchased, and respecting data quality in a multidimensional feature space.

Explore the types of variables, distinguishing dependent from independent variables, and see how observed values like age, gender, zip code, and purchase counts inform predictive modeling.

Compare independent and dependent variables, noting that the independent variable can be manipulated to explain effects on the dependent variable, while gender cannot, with examples from age brackets and generations.

Identify extraneous variables beyond price, including control, moderating, and intervening types; control keeps price constant, moderating relates to returns, and intervening infers unquantified effects on customer behavior.

Explore 13 predictive modeling algorithms, including time series, regression, association, clustering, outlier detection, decision trees, neural networks, Naive Bayes, SVM, uplift, and survival analysis, to forecast data and reveal trends.

Compare qualitative and quantitative forecasting methods, using data mining and statistical analysis to identify trends and predict future events, with qualitative relying on expert judgment for new products and technology.

Time series uses historical data at regular intervals to forecast future values and reveal trend, cyclical, seasonal, and irregular patterns, with smoothing methods like moving averages.

Master smoothing for time series with moving averages and weighted moving averages, plus single, double (Holt's), and triple (Winter's) exponential smoothing, using alpha and past values to forecast.

Explore double exponential smoothing and the trend smoothing constant beta for time series forecasting, then survey five regression algorithms from linear to multiple linear regression.

Explore clustering algorithms that group unlabeled data by similarity or descriptive concepts, using distance-based, exclusive, overlapping, hierarchical, and probabilistic approaches, including k-means, fuzzy c-means, and mixtures of gaussians.

Explore fuzzy C means clustering with degrees of membership and iterative center updates, and note hierarchical clustering, Gaussian mixture models, decision trees, and outlier detection.

Explore neural networks and learning models, including Kohonen self-organizing maps, Hopfield nets, bump tree network, Monte Carlo analysis, factor analysis, and Naive Bayes theorem.

Learn support vector machines and uplift modeling to predict treatment effects and customer behavior, then apply survival analysis and Bayes theorem to time-to-event outcomes.

Explore econometrics-focused finance modeling using Eviews, with hands-on exploration of descriptive statistics, correlogram, and cointegration test, plus regression, autocorrelation, and arch models.

Navigate the Eviews GUI, start the software, and import foreign data from formats like WF1, DBF, Excel, SAS, SPSS, and text; Minitab data isn't supported.

Explore the Eviews GUI for estimating regression equations and viewing outputs like R-squared and Durbin-Watson. Generate returns, create volatility graphs, and perform descriptive statistics.

Generate and interpret log returns and descriptive statistics, including standard deviation, with a t test in Eviews across five mutual fund data sets to explore econometrics in financial markets.

Generate descriptive statistics for fund returns using Eviews, including mean, median, maximum, standard deviation, and Jarque-Bera; interpret kurtosis and volatility to assess risk.

Learn to interpret descriptive statistics and their investment significance. Use Jarque-Bera and standard deviation to assess volatility and risk, illustrated by HDFC Equity Fund and HDFC Mid Cap Opportunities Fund.

Explore volatility analysis by generating spike-based volatility graphs, interpreting Jarque-Bera and standard deviation to compare fund risk, and linking descriptive statistics to econometric data interpretation.

Apply descriptive analysis to stock indices by generating log returns and comparing close prices of BSE Sensex, mid cap, and small cap using Eviews.

Generate log returns for large cap, mid cap, and small cap indices in Eviews, then compute descriptive statistics and compare volatility across groups.

Analyze foreign-exchange data from AUD, GBP, and euro, computing log returns and generating descriptive statistics and volatility graphs. Note Brexit-driven moves and their impact on Jarque-Bera and kurtosis.

Explore how the Brexit effect drives GBP and euro volatility and skewness, shown through descriptives and volatility graphs, with notes on correlation and regression modeling.

Define correlation as the relationship indicator between two variables. Show that r lies between -1 and 1 with signs for positive, negative, zero correlation.

Learn to generate a correlation matrix in EViews using log returns for five funds, then interpret the results to assess diversification and investment viability.

Analyze the correlation matrix of sectoral and non-sectoral mutual funds to identify high correlations within sector funds and low or negative links with elss, diversified equity, and mid-cap funds.

Learn to create scatter plots in Eviews to visualize positive, negative, and zero correlations, interpret regression lines, and align visuals with the correlation matrix.

Analyze the correlation of Sensex stocks by computing a three by three correlation matrix from log returns of Reliance, Infosys, and the BSE Sensex, to identify the best investment script.

Explore scatter plots and volatility graphs to compare Reliance, Infosys, and the BSE Sensex, identify weak and strong positive correlations, and view regression lines for insight.

Generate a correlation matrix for a multi-asset price dataset, calculate logarithmic returns, and interpret relationships among gold, natural gas, and Swiss franc to inform hedge and portfolio decisions.

Analyze how to generate log returns and interpret a correlation matrix to build risk-aware portfolio combinations, using gold, Swiss franc, gas, and Nifty assets.

Analyze correlation interpretations to guide risk management, diversification, and hedging decisions across gold, Swiss franc, and Nifty via derivatives.

Explore scatter plots and a scatter plot matrix of multi-asset returns, including Nifty, gold, natural gas, and Swiss franc, revealing near unity correlations and regression lines.

Explore volatility patterns using scatter plots and volatility graphs, highlighting Swiss franc and natural gas spikes. Assess the relative safety of the Nifty and learn to present these insights clearly.

Learn linear regression basics—y equals mx plus c—with x as predictor and y as response, including simple, multiple, logistic, and polynomial forms and stats like t, p, and r squared.

Explore regression analysis for generating returns from stock data, detailing the dependent variable, constants, coefficients, and outputs such as p-values, r-squared, f-statistics, and Akaike Schwarz Hannan-Quinn criteria.

Learn to read the regression equation and estimation output, and interpret t-statistics, p-values, and R-squared to assess predictor significance in stock returns, reliance, and the BSE Sensex.

Analyze descriptive statistics and volatility patterns across the BSE Sensex, midcap, and small cap indices, using mean, standard deviation, and Jarque-bera to inform investment decisions.

Generate and interpret an ordinary least squares estimation output for Tata Motors log returns against Sensex movements using Eviews, including regression equation, t statistics, and R-squared insights.

Explain the interpretation of the regression output and the significance of coefficients, compare the correlation graphs for Tata Motors and Sensex, and discuss the r-squared and volatility factors.

Analyze volatility scatter plots and regression lines to compare stock volatility, like reliance and Tata motors, using Eviews, with emphasis on interpreting correlation in a one-year financial data context.

Analyze estimation outputs and graphs for a mutual fund case study, comparing SBI Pharma Fund to the BSc healthcare index using log returns and least squares regression.

Interpret estimation outputs to reveal insignificant predictors, near-zero r-squared, and flat regression plots, guiding judgment on best-fit models in predictive analytics using R, Minitab, SPSS, and SAS.

Analyze volatility graphs showing spikes from news events like Fed rate delays, Brexit, and Chinese market swings, and explain why models are not a best fit for predicting fund movements.

Explore how the GBP and euro influence the Australian dollar through correlation matrices and regression equations, using log returns and EViews to compare bidirectional effects.

Explore regression-based estimation outputs for aud, gbp, and euro, detailing coefficients and how gdp and other factors influence currency values.

Investigate how the Australian dollar, euro, and pound sterling interact through regression equations and estimation outputs, using scatter plots and regression lines to compare reserve currency roles against US dollars.

Clean the data by removing observations and reimporting, then run a regression with p/e and peg and growth as variables. Peg and growth are insignificant, with low r-squared.

Study how Swiss franc and natural gas influence gold through regression analysis and log returns, and interpret the estimation output in Eviews with a modest 6% R-squared.

Analyze scatterplot matrices and regression outputs to explore relationships among gold, Swiss franc, and natural gas. Swiss franc is the only significant independent variable predicting gold, with r square 6%.

Estimate a multiple regression of gold prices on Singapore Nifty, natural gas, and Swiss franc, updating equation and noting that Swiss franc significantly affects gold while others show limited impact.