Advanced Data Science Techniques in SPSS
3.8 (88 ratings)
Course Ratings are calculated from individual students’ ratings and a variety of other signals, like age of rating and reliability, to ensure that they reflect course quality fairly and accurately.
9,803 students enrolled

Advanced Data Science Techniques in SPSS

Hone your SPSS skills to perfection - grasp the most high level data analysis methods available in the SPSS program.
3.8 (88 ratings)
Course Ratings are calculated from individual students’ ratings and a variety of other signals, like age of rating and reliability, to ensure that they reflect course quality fairly and accurately.
9,803 students enrolled
Created by Bogdan Anastasiei
Last updated 2/2018
English
English [Auto]
Current price: $55.99 Original price: $79.99 Discount: 30% off
23 hours left at this price!
30-Day Money-Back Guarantee
This course includes
  • 6.5 hours on-demand video
  • 9 articles
  • 8 downloadable resources
  • Full lifetime access
  • Access on mobile and TV
  • Certificate of Completion
Training 5 or more people?

Get your team access to 4,000+ top Udemy courses anytime, anywhere.

Try Udemy for Business
What you'll learn
  • Perform advanced linear regression using predictor selection techniques
  • Perform any type of nonlinear regression analysis
  • Make predictions using the k nearest neighbor (KNN) technique
  • Use binary (CART) trees for prediction (both regression and classification trees)
  • Use non-binary (CHAID) trees for prediction (both regression and classification trees)
  • Build and train a multilayer perceptron (MLP)
  • Build and train a radial basis funcion (RBF) neural network
  • Perform a two-way cluster analysis
  • Run a survival analysis using the Kaplan-Meier method
  • Run a survival analysis using the Cox regression
  • Validate the predictive techniques (KNN, trees, neural networks) using the validation set approach and the cross-validation
  • Save a predictive analysis model and use it for predictions on future new data
Course content
Expand all 87 lectures 06:41:24
+ Getting Started
1 lecture 05:16

What you are going to learn in this course, exactly.

Preview 05:16
+ Advanced Linear Regression Techniques
11 lectures 43:48

Here we will study the techniques and criteria available in SPSS to enter or to remove a predictor variable from a regression analysis.

Preview 05:34

The data set used for our practical example of stepwise regression, as well as our research goal.

Our Practical Example
02:26

How to execute the stepwise regression in SPSS using the Stepwise technique.

Executing the Stepwise Regression Method
03:04

How to read the output tables of a stepwise regression method - and how to detect the best regression model.

Interpreting the Results of the Stepwise Method
09:12

How to perform a stepwise regression in SPSS using the Forward method.

Executing the Forward Selection Regression
01:14

How to interpret the output tables of the Forward method, and how to identify the model with the highest prediction accuracy.

Interpreting the Results of the Forward Selection Method
04:51

How to run the stepwise regression in SPSS using the Backward selection method.

Executing the Backward Selection Regression
00:53

How to interpret the output of the Backward selection method, and how to identify the best model.

Interpreting the Results of the Backward Selection Method
04:37

What are nested models and how we can use the Remove method in SPSS in order to compare their prediction accuracy and seelct the most effective model.

Comparing Nested Models Using the Remove Method
04:46

How to enter the predictors in order to define our nested models in SPSS.

Executing the Regression Analysis with the Remove Method
03:35

How to detect the model with the greatest prediction accuracy out of two or more nested regression models.

Interpreting the Results of the Remove Method
03:36
+ Nonlinear Regression Analysis
10 lectures 50:53

Here I remind you the most common types of nonlinear functions you will find in practice. Most of them will be approached in this section.

Preview 03:27

The nonlinear relationships can be divided into two groups: linearizable and non-linearizable. This is important, because the linearizable relationships can be modeled in two ways in SPSS. So let's see what relationships we can found in each group.

Preview 04:41

In this lecture we are going to run a quadratic regression using the linear regression module in SPSS. In other words, we are going to linearize the quadratic equation before performing the analysis.

Performing a Quadratic Regression in SPSS (1)
06:51

How to perform a quadratic regression, this tyme using the nonlinear regression module in SPSS (so without linearizing the equation).

Performing a Quadratic Regression in SPSS (2)
06:55

How to execute a cubic regression by reducing the cubic equation to linearity (so we are going to use the linear regression module in SPSS).

Performing a Cubic Regression in SPSS (1)
07:02

How to run a cubic regression using the nonlinear regression module (without reducing the equation to linearity).

Performing a Cubic Regression in SPSS (2)
04:11

How to run an inverse regression in SPSS by reducing the inverse function to linearity.

Performing an Inverse Regression in SPSS (1)
04:22

How to execute an inverse regression without reducing the equation to linearity, using the nonlinear regression module in SPSS.

Performing an Inverse Regression in SPSS (2)
02:54

How to run a nonlinear regression when the relationship between the dependent variable and the predictor is deemed to be exponential (and, of course, how to interpret the results).

Performing a Nonlinear Regression With an Exponential Relationship
04:45

How to run a nonlinear regression when the relationship between the variables can be described by a logistic function.

Performing a Nonlinear Regression With a Logistic Relationship
05:45
+ K Nearest Neighbor in SPSS
8 lectures 43:31

We explain how the KNN method works (using a graphicla representation), when it can be used.

Preview 03:37

The prediction accuracy of the KNN technique depends on the number of neighbors (K). Here we show how to find out the optimal value of K using two validation techniques: validation set approach and cross-validation.

Selecting the Optimal Number of Neighbors
04:13

The data set and the practical problem we are going to use in order to illustrate how to execute the KNN method.

Our Practical Example
01:48

How to execute the KNN method in SPSS, step by step. For this first example we will set a fixed value for K (the number of neighbors) and use the validation set approach to validate our KNN model.

Performing the KNN technique
05:08

In this lecture we explain how to interpret all the tables and charts of the KNN technique (I remind you, the validation method used was the validation set approach). Most important, we will assess the prediction accuracy of our KNN model.

Interpreting the results of the KNN analysis
17:33

Here we execute the KNN technique in SPSS, this time using the cross-validation to determine the optimal number of neighbors (K will take values between 1 and 100).

Finding the Optimal Number of Neighbors with Cross-Validation
01:55

So what is the optimal number of neighbors in our concrete situation? In this lecture we'll find out!

Interpreting the Cross-Validation Results
01:45

How to save your KNN model and use it later, for predictions on new data.

Using the KNN Model for Future Predictions
07:32
+ Introduction to Decision Trees
4 lectures 18:51

Here we will familiarize with the concept of decision trees as it is used in the predictive analysis. We will talk about tree structure, splitting rules/criteria, types of nodes etc.

Preview 06:21

We introduce the first major type of decision trees, Classification and Regression Trees (CART). In this type we only find binary trees, where each node can be split into two sub-nodes only.

Binary Trees (CART)
05:48

We present the second important type of decision trees, Chi-Square Automatic Interaction Detector (CHAID) trees. In this type we find non-binary trees, where each node can be split into more than two sub-nodes.

Non-Binary Trees (CHAID)
05:19

It is useful to know both the advantages, as well as the weak points of the decision trees in predictive analysis.

Advantages and Disadvantages of Decision Trees
01:23
+ Growing Binary Trees (CART) in SPSS
11 lectures 01:03:59

How to define a binary regression tree (when the dependent variable is numeric). The validation method used will be the validation set approach.

Growing a Binary Regression Tree (CART)
06:11

Know how to "read" a tree. So here we present all the information a CART regression tree contains. Based on this information you can use the tree for predictions.

Intepreting a Binary Regression Tree (1)
09:10

How to interpret the other output provided by SPSS: the tree in table format, the predictor importance chart and the risk (this risk is directly related to the tree prediction accuracy).

Intepreting a Binary Regression Tree (2)
04:25

Since the SPSS program does not automatically compute the R squared of our tree, so we have to compute it manually. in this lecture I will show you how.

Computing the R Squared
02:57

How to build a CART regression tree, this time using cross-validation in order to find the tree with the best prediction accuracy.

Growing a CART Regression Tree with Cross-Validation
02:07

How to read the output for our CART regression tree built using the cross-validation method.

Interpreting the Cross-Validation Results for a Regression Tree
02:47

How to build a CART tree where the dependent variable is categorical - step by step.

Growing a CART Classification Tree in SPSS
05:18

How to "read" a CART classification tree: all the information contained in the tree, and how to make predictions using the tree.

Interpreting the CART Classification Tree
13:01

In this lecture we are going to build a binary classification tree using the cross-validation method to validate it.

Growing a CART Classification Tree with Cross-Validation
02:13

How to "read" the tree and make predictions based on it.

Interpreting the Cross-Validation Results for a Classification Tree
03:59

How to save your binary tree and use it later, for predictions on new data.

Using Binary Trees for Future Predictions
11:51
+ Growing Non-Binary Trees (CHAID) in SPSS
7 lectures 39:09

How to execute the SPSS procedure for growing a CHAID tree when the response variable is continuous (numeric).

Building a CHAID Regression Tree
03:21

How to read the information in the tree and how to make predictions about the value of the dependent variable.

Interpreting a CHAID Regression Tree
07:49

How to build a CHAID tree using the cross-validation technique to validate it (and interpret the output).

Growing a CHAID Regression Tree with Cross-Validation
03:45

How to execute the SPSS procedure for growing a CHAID tree when the response variable is categorical.

Building a CHAID Classification Tree
04:33

How to make prediction based on the non-binary classification tree.

Interpreting a CHAID Classification Tree
08:02

How to execute the SPSS procedure for building a non-binary classification tree using the cross-validation (and interpret the results).

Growing a CHAID Classification Tree with Cross-Validation
04:33

How to save your non-

binary tree and use it later, for predictions on new data.
Using Non-Binary Trees for Future Predictions
07:06
+ Introduction to Neural Networks
4 lectures 14:32

Here we will learn about nodes and layers, as well as about the two main types aof neural networks.

Preview 03:48

In this lecture we'll learn about how a neuron (node) in a neural network processes the information using two operations: summation and activation.

Preview 03:25

Now it's time to learn about the activation functions that are most widely used in neural networks.

Activation Functions
03:40

How a neural network learns; i.e. how the information is transmitted form the input nodes to the output nodes, and how the path weights are adjusted in order to minimize the prediction error.

Neural Network Learning Process
03:39
+ Training a Multilayer Perceptron (MLP) in SPSS
4 lectures 24:51

The procedure for creating a MLP in SPSS, step by step, with all the explanations.

Building a Multilayer Perceptron
06:42

How to read the output for a MLP, and how to evaluate the prediction accuracy.

Interpreting the Multilayer Perceptron
09:28

In this lecture we'll learn how to interpret the ROC curve related to our MLP (ROC stands for Receiver Operating Characteristics). This curve summarizes the performance of a neural network.

Interpreting the ROC Curve
04:18

How to save your MLP

and use it later, for predictions on new data.
Using the Multilayer Perceptron for Future Predictions
04:23
+ Training a Radial Basis Function (RBF) Neural Network in SPSS
3 lectures 13:54

The SPSS procedure for creating an RBF network, step by step.

Building an RBF Neural Network
05:08

How to interpret the output of the analysis and how to assess the network prediction accuracy.

Interpreting the RBF Network
05:38

How to save your RBF network and use it later, for predictions on new data.

Using the RBF Network for Future Predictions
03:08
Requirements
  • SPSS program installed (version 21+)
  • Basic SPSS knowledge
  • Basic or intermediate statistics knowledge
Description

Become a Top Performing Data Analyst – Take This Advanced Data Science Course in SPSS!

Within a few days only you can master some of the most complex data analysis techniques available in the SPSS program. Even if you are not a professional mathematician or statistician, you will understood these techniques perfectly and will be able to apply them in practical, real life situations.

These methods are used every day by data scientists and data miners to make accurate predictions using their raw data. If you want to be a high skilled analyst, you must know them!

Without further ado, let’s see what you are going to learn…

  • Stepwise regression analysis, a technique that helps you select the best subset of predictors for a regression analysis, when you have a big number of predictors. This way you can create regression models that are both parsimonious and effective.
  • Nonlinear regression analysis. After finishing this course, you will be able to fit any nonlinear regression model using SPSS.
  • K nearest neighbor, a very popular predictive technique used mostly for classification purposes. So you will learn how to predict the values of a categorical variable with this method.
  • Decision trees. We will approach both binary (CART) and non-binary (CHAID) trees. For each of these two types we will consider two cases: the case of response dependent variables (regression trees) and the case of categorical response variables (classification trees).
  • Neural networks. Artificial neural networks are hot now, since they are a suitable predictive tool in many situations. In SPSS we can train two types of neural network: the multilayer perceptron (MLP) and the radial basis function (RBF) network. We are going to study both of them in detail.
  • Two-step cluster analysis, an effective grouping procedure that allows us to identify homogeneous groups in our population. It is useful in very many fields like marketing research, medicine (gene research, for example), biology, computer science, social science etc.
  • Survival analysis. If you have to estimate one of the following: the probable time until a certain event happens, what percentage of your population will suffer the event or which particular circumstances influence the probability that the event happens, than you need to apply on of the survival analysis method studied here: Kaplan-Meier or Cox regression.

For each analysis technique, a short theoretical introduction is provided, in order to familiarize the reader with the fundamental notions and concepts related to that technique. Afterwards, the analysis is executed on a real-life data set and the output is thoroughly explained.

Moreover, for some techniques (KNN, decision trees, neural networks) you will also learn:

  • How to validate your model on an independent data set, using the validation set approach or the cross-validation
  • How to save the model and use it for make predictions on new data that may be available in the future.

Join right away and start building sophisticated, in-demand data analysis skills in SPSS!

 

 

Who this course is for:
  • students
  • PhD candidates
  • academic researchers
  • business researchers
  • University teachers
  • anyone who is passionate about data analysis and data science