Predictive Modeling Training Course Content

Spread the love

Predictive Modeling Training Course Content Details

What is Predictive Modeling

Predictive modeling uses statistics to predict outcomes. most frequently the event one desires to predict is within the future, however prophetical modelling is applied to any variety of unknown event, no matter once it occurred. for instance, prophetical models ar typically wont to find crimes and determine suspects, once the crime has taken place.

In several cases the model is chosen on the idea of detection theory to undertake to guess the chance of AN outcome given a collection quantity of computer file, {for example|for instance|as AN example} given an email determinative however seemingly that it’s spam.

Models will use one or a lot of classifiers in attempting to work out the chance of a collection of knowledgehappiness to a different set, say spam or ‘ham’.

Depending on definitional boundaries, prophetical modelling is substitutable with, or mostly overlapping with, the sphere of machine learning, because it is a lot of ordinarily stated in educational or analysis and development contexts. once deployed commercially, prophetical modelling is commonly stated as prophetical analytics. for more info click here

Predictive Modeling Course Content

01.. Introduction

  • What is Data Mining?
  • Terminology of Data Mining
  • Types of Variables: Interval, Nominal (Unordered Categorical), and Ordinal
  • The Distinct Purposes of Hypothesis Testing versus Prediction (Read Breiman article)
  • Data Mining from a Process Perspective
  • Data Mining Methods Classified by Nature of the Data

02. Overview of the Data Mining Process

  • Core Ideas in Data Mining
  • Classification
  • Prediction
  • Association Rules
  • Data Reduction
  • Data Exploration

03. Data Exploration and Data Refinement

  • Data Summaries
  • Data Visualization
  • Treatment of Missing Observations
  • Detection of Outliers – the Box Plot
  • Correlation Analysis

04. Variable Importance and Dimension Reduction

  • Binning: Reducing the Number of Categories in Categorical Variables
  • Principal Component Analysis of Continuous Variables
  • Dimension Reduction using Best Subset Regression and LASSO Modelling Techniques
  • Dimension Reduction using Bivariate Association Probabilities

05. Evaluation Methods for Prediction and Classification Problems

  • Prediction Measures: MAE, MSE, RMSE, MAPE, MSPE, and RMSPE
  • Application to Validation and Test Data Sets
  • Avoiding Overtraining

06. Prediction Methods

  • Linear Regression: Best Subset Selection
  • Forward Selection
  • Backward Selection
  • Step-wise Regression
  • All Subsets Regression
  • Data Visualization

07. Classification Methods

  • The Naïve Rule
  • Naïve-Bayes Classifier
  • K-Nearest Neighbors
  • Classification Trees
  • Neural Nets
  • Logistic Regression
  • Support Vector Machines (SVM)

08. Ensemble Methods

  • Nelson and Granger-Ramanathan Methods for Continuous Targets
  • Majority Voting for Categorical
  • Targets
  • Bagging
  • Boosting

Predictive Modeling Training

Sharing is caring!

Updated: May 18, 2017 — 12:14 pm

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.