Machine Learning

Neural Network

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Predict Data Point
  • Optimization Method
  • Number of Hidden Neurons
  • Activation Function
  • Model Name

Optional Parameters:

  • Max Iterations

Description:

The Train Neural Network action trains a Neural Network machine learning model which can be used later for predictive tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a Neural Network Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

Note - We will be using the Iris Flower Species Sample Dataset as our Working Data collection.

iris flower species sample dataset

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Select an Optimization Method:

  • Parallel Resilient Backpropagation
  • Levenberg Marquadt

4

Enter a Number of Hidden Neurons.

5

Set a number for Max Iterations (optional).

6

Select an Activation Function:

  • Bipolar Sigmoid Function
  • Sigmoid Function
  • Identity Function
  • Rectified Linear Function
  • Linear Function
  • Threshold Function

7

Select the Predict Data Point (Response Feature).

8

Enter a Model Name.

neural-network-how-to-1

9

Click OK.

Running the action...

neural-network-how-to-2

Decision Tree

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Predict Data Point
  • Model Name

Description:

The Train Decision Tree action trains a ID3 Decision Tree machine learning model which can be used later for predictive tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a Decision Tree Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Select the Predict Data Point (Response Feature).

4

Enter a Model Name.

5

Click OK.

K-Means Cluster

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Number of Clusters
  • Distance Function
  • Model Name

Optional Parameters:

  • Max Iterations
  • Tolerance
  • Compute Covariance
  • Compute Error

Description:

The Train K-Means Cluster action trains a K-Means Cluster machine learning model which can be used for unsupervised learning tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a K-Means Cluster Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

Note - We will be using the Iris Flower Species Sample Dataset as our Working Data collection.

iris flower species sample dataset

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Set the Number of Clusters (K).

4

Set a number for Max Iterations (optional).

5

Set a number for the Tolerance (optional).

6

Check Compute Covariance to compute the covariance (optional).

7

Check Compute Error to compute the error (optional).

8

Select an Distance Function:

  • Squared Euclidean Distance

9

Enter a Model Name.

k-means-clustering-how-to-1

10

Click OK.

Running the action...

k-means-clustering-how-to-2

K-Nearest Neighbors

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Predict Data Point
  • K
  • Model Name

Description:

The Train K-Nearest Neighbors action trains a K-Nearest Neighbors machine learning model which can be used for unsupervised learning tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a K-Nearest Neighbors Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Select the Predict Data Point (Response Feature).

4

Set a number for Nearest Neighbors (K).

5

Enter a Model Name.

6

Click OK.

Linear Regression

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Predict Data Point
  • Model Name

Description:

The Train Linear Regression action trains a Linear Regression machine learning model which can be used later for quantitative predictive tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a Linear Regression Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

Note - We will be using the Product Sales Sample Dataset as our Working Data collection.

product sales sample dataset

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Select the Predict Data Point (Response Feature).

4

Enter a Model Name.

linear-regression-how-to-1

5

Click OK.

Running the action...

linear-regression-how-to-2

Logistic Regression

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Predict Data Point
  • Model Name

Optional Parameters:

  • Max Iterations
  • Tolerance

Description:

The Train Logistic Regression action trains a Logistic Regression machine learning model which can be used later for qualitative predictive tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a Logistic Regression Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Select the Predict Data Point (Response Feature).

4

Set a number for Max Iterations (optional).

5

Set a number for the Tolerance (optional).

6

Enter a Model Name.

7

Click OK.

Naive Bayes

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Predict Data Point
  • Model Name

Description:

The Train Naive Bayes action trains a Naive Bayes machine learning model which can be used later for predictive tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a Naive Bayes Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Select the Predict Data Point (Response Feature).

4

Enter a Model Name.

5

Click OK.

Min Mean Distance

Required Parameters:

  • Working Data
  • Feature Data Point(s) (at least 1)
  • Predict Data Point
  • Distance Function
  • Model Name

Description:

The Train Min Mean Distance action trains a Min Mean Distance machine learning model which can be used later for classification tasks. The model is learned during workflow execution and then saved to the cloud.

How to Train a Min Mean Distance Model
Step Description

1

Select the Working Data collection that contains the Feature Data Point(s).

2

Select a Data Point and click the green (+) to add a Feature to the model.

3

Select the Predict Data Point (Response Feature).

4

Select an Distance Function:

  • ArgMax
  • Bhattacharyya
  • BrayCurtis
  • Canberra
  • Chebyshev
  • Cosine
  • Dice
  • Euclidean
  • Hamming
  • Hellinger
  • Jaccard
  • Kulczynski
  • Levenshtein
  • Manhalanobis
  • Manhattan
  • Minkowski
  • Modular
  • SquareEuclidean
  • SquareManhabolis

5

Enter a Model Name.

6

Click OK.