SAS-A00-255 Predictive Modeling Use SAS EM

SASPredictive Modeling Use SAS EM考试主要考察深度学习和机器学习理论知识的理解，以及对SAS代码和软件的掌握程度。我从前期准备、需要学习的内容、现场考试这几个方面说下如何拿到EM认证。

目录

 

---

1、前期准备

1.1 确定考点大纲

到官网找最新的考试大纲。大纲分为4个部分，分别是对数据源的操作、建模、模型评估和模式分析。

During the testing of these objectives; you will be expected to perform common tasks, such as:

 Create a new project in Enterprise Miner
 Open an existing project in Enterprise Miner
 Add diagrams to projects in Enterprise Miner
 Create libraries within Enterprise Miner
 Add nodes to diagrams in Enterprise Miner
 Copy nodes within Enterprise Miner
 Connect nodes to create process flows in Enterprise Miner
 Change interactive sampling methods for data exploration
 Work with the Help functionality within Enterprise Miner

1）Data Sources - 20-25%

Create data sources from SAS tables in Enterprise Miner

 Use the Basic Metadata Advisor
 Use the Advanced Metadata Advisor
 Customize the Advanced Metadata Advisor
 Set Role and Level meta data for data source variables
 Set the Role of the table (raw, scoring, transactional, etc)

Explore and assess data sources

 Create and interpret plots, including Histograms, Pie charts, Scatter plot, Time series, Box plot
 Identify distributions
 Find outlying observations
 Find number (or percent) of missing observations
 Find levels of nominal variables
 Explore associations between variables using plots by highlighting and selecting data
 Compare balanced and actual response rates when oversampling has been performed
 Explore data with the STAT EXPLORER node.
 Explore input variable sample statistics
 Browse data set observations (cases)

Modify source data

 Replace zero values with missing indicators using the REPLACEMENT node
 Use the TRANFORMATION node to be able to correct problems with input data sources, such as variable distribution or outliers.

 Use the IMPUTE node to impute missing values and create missing value indicators
 Reduce the levels of a categorical variable
 Use the FILTER node to remove cases

Prepare data to be submitted to a predictive model

 Select a portion of a data set using the SAMPLE node
 Partition data with the PARTITION Node
 Use the VARIABLE SELECTION node to identify important variables to be included in a predictive model.
 Use the PARTIAL LEAST SQUARES node to identify important variables to be included in a predictive model.
 Use a DECISION TREE or REGRESSION nodes to identify important variables to be included in a predictive model.

 

2）Building Predictive Models - 35-40%

Describe key predictive modeling terms and concepts

 Data partitioning: training, validation, test data sets
 Observations (cases), independent (input) variables, dependent (target) variables
 Measurement scales: Interval, ordinal, nominal (categorical), binary variables
 Prediction types: decisions, rankings, estimates
 Dimensionality, redundancy, irrelevancy
 Decision trees, neural networks, regression models
 Model optimization, overfitting, underfitting, model selection
 Describe ensemble models

Build predictive models using decision trees

 Explain how decision trees identify split points
 Build decision trees in interactive m