For 22.2 - Configuring and Training CIC Advisor Models with Primavera Cloud Data

To manage information displayed in CIC Advisor, as administrators you have to select models that need to be trained for predicting outcomes.

R2: It measures the proximity of the data points to the fitted regression line. The higher the R2 score, the better. It is calculated as: Explained Variation / Total Variation.Note: A seed model or initial model is also delivered for each problem. Oracle recommends that you train the seed model before using it for predictions.

To set up and train models with specific features enabled:

Sign in to the administration application of CIC Advisor.
http://<host>:<port>/cicadmin
In the sidebar, select ML Workbench.
From the Data Source list, select a Primavera Cloud data source.
From the Problem list, select a predefined area of improvement that you want to focus in CIC Advisor for the selected Primavera Cloud data source.
In the Model section, add a model for the selected problem as follows:
1. Select Add Model, and then select OK to confirm your selection.
  By default, the model is named as CustomModel1.
  
  The following information is displayed for each model:
  - Model Name: A user-friendly name to identify a model.
  - Prediction Enabled: Indicates that a model is used for predictions. For each problem, only one model which has been trained can be used for predictions.
  - Training Enabled: Indicates that a model is being trained. Multiple models can be trained for a problem.
  - Select Group:
  - Accuracy: It is the ratio of correct predictions to the total number of predictions.
  - Precision: It is a measure of correctness on positive predictions, and a measure of how many are actually delayed out of all the activities that are predicted to be delayed.
  - Recall: It is a measure of a model's ability to find all the positive results actually in the data. Of all the activities that are actually delayed, how many the model correctly identified.
  - MAE: The Mean Absolute Error (MAE) measures the average magnitude of the error in the set of predictions without considering the direction. The lower the score, the better.
  - MAPE: The mean absolute percentage error (MAPE) measures the prediction accuracy of a forecasting method. It is expressed in terms of a percentage value.
  - R2: It measures the proximity of the data points to the fitted regression line. The higher the R2 score, the better. It is calculated as: Explained Variation / Total Variation.
  - RMSE: The Root Mean Square Error (RMSE) measures the average magnitude of the error. It's the square root of the average of the squared difference between predictions and actual observations. The lower the score, the better.
  - Create Date: The date on which the model was added.
  - Update Date: The date on which the model was previously updated.
  - Train Status: The current status of the model being trained. Choices include: Failed, Completed.
  - Train Start Time: Time when the training process for the model was initiated.
  - Train Finish Time: Time when the training process for the model was completed.
  - Last Training Date: The date on which the model previously trained.
    Note: Prediction can be enabled for a model only if a model has a Last Training Date.
  - Training Log: Click the View Logs link to view the log file containing details corresponding to each model that has been trained.
  - Prediction Status: The current status of the model enabled for prediction. Choices include: Failed, Completed.
  - Prediction Start Time: Time when the prediction process for the model was initiated.
  - Prediction Finish Time: Time when the prediction process for the model was completed.
  - Prediction Log: Click the View Logs link to view details of the prediction process run for the model that has been enabled for prediction.
2. In the Model Name field, select Edit, and rename the model.
3. In the Description section, enter information that describes the purpose of the model.
Enable features for a model:
1. In the Model section, select a row to select the model whose feature you want to enable.
2. In the Features section, select the Enabled check box to activate each feature for the selected model.
3. Select Save.
To add multiple models for each problem, repeat steps 5 and 6 for each problem.
To train the models:
1. In the Model section, select the Training Enabled check box for one or more models across all problems.
2. Select Train Model(s).

Related Topics