Schulung - IBM 0A018G - Data science without a Ph.D. Using IBM SPSS Modeler (v18.1.1)

Course Outline

1:  Introduction to data science and IBM SPSS Modeler
    •  Explain the stages in a data-science project, using the CRISP-DM methodology
    •  Create IBM SPSS Modeler streams
    •  Build and apply a machine learning model
2:  Setting measurement levels
    •  Explain the concept of field measurement level
    •  Explain the consequences of incorrect measurement levels
    •  Modify a field's measurement level
3:  Exploring the data
    •  Audit the data
    •  Check for invalid values
    •  Take action for invalid values
    •  Impute missing values
    •  Replace outliers and extremes
4:  Using automated data preparation
    •  Automatically exclude low quality fields
    •  Automatically replace missing values
    •  Automatically replace outliers and extremes
5:  Partitioning the data
    •  Explain the rationale for partitioning the data
    •  Partition the data into a training set and testing set
6:  Selecting predictors
    •  Automatically select important predictors (features) to predict a target
    •  Explain the limitations of automatically selecting features
7:  Using automated modeling
    •  Find the best model for categorical targets
    •  Find the best model for continuous targets
    •  Explain what an ensemble model is
8:  Evaluating models
    •  Evaluate models for categorical targets
    •  Evaluate models for continuous targets
9:  Deploying models
    •  List two ways to deploy models
    •  Export scored data

Objective

• Introduction to data science and IBM SPSS Modeler
• Setting measurement levels
• Exploring the data
• Using automated data preparation
• Partitioning the data
• Selecting predictors
• Using automated modeling
• Evaluating models
• Deploying models