Training the pipeline

The train mode, gds.alpha.pipeline.nodeRegression.train, is responsible for data splitting, feature extraction, model selection, training and storing a model for future use. Running this mode results in a regression model of type NodeRegression, which is then stored in the model catalog. The regression model can be applied on a graph to predict property values for new nodes.

More precisely, the training proceeds as follows:

  1. Apply nodeLabels and relationshipType filters to the graph.

  2. Apply the node property steps, added according to Adding node properties, on the whole graph.

  3. Select node properties to be used as features, as specified in Adding features.

  4. Split the input graph into two parts: the train graph and the test graph. This is described in Configuring the node splits. These graphs are internally managed and exist only for the duration of the training.

  5. Split the nodes in the train graph using stratified k-fold cross-validation. The number of folds k can be configured as described in Configuring the node splits.

  6. Each model candidate defined in the parameter space is trained on each train set and evaluated on the respective validation set for every fold. The evaluation uses the specified primary metric.

  7. Choose the best performing model according to the highest average score for the primary metric.

  8. Retrain the winning model on the entire train graph.

  9. Evaluate the performance of the winning model on the whole train graph as well as the test graph.

  10. Retrain the winning model on the entire original graph.

  11. Register the winning model in the Model Catalog.

The above steps describe what the procedure does logically. The actual steps as well as their ordering in the implementation may differ.
A step can only use node properties that are already present in the input graph or produced by steps, which were added before.

1. Metrics

The Node Regression model in the Neo4j GDS library supports the following evaluation metrics:

  • MEAN_SQUARED_ERROR

  • ROOT_MEAN_SQUARED_ERROR

  • MEAN_ABSOLUTE_ERROR

More than one metric can be specified during training but only the first specified — the primary one — is used for evaluation, the results of all are present in the train results.

2. Syntax

Run Node Regression in train mode on a named graph:
CALL gds.alpha.pipeline.nodeRegression.train(
  graphName: String,
  configuration: Map
) YIELD
  trainMillis: Integer,
  modelInfo: Map,
  modelSelectionStats: Map,
  configuration: Map
Table 1. Parameters
Name Type Default Optional Description

graphName

String

n/a

no

The name of a graph stored in the catalog.

configuration

Map

{}

yes

Configuration for algorithm-specifics and/or graph filtering.

Table 2. Configuration
Name Type Default Optional Description

pipeline

String

n/a

no

The name of the pipeline to execute.

nodeLabels

List of String

['*']

yes

Filter the named graph using the given node labels.

relationshipTypes

List of String

['*']

yes

Filter the named graph using the given relationship types.

concurrency

Integer

4

yes

The number of concurrent threads used for running the algorithm.

targetProperty

String

n/a

no

The target property of the node. Must be of type Integer or Float.

metrics

List of String

n/a

no

Metrics used to evaluate the models.

randomSeed

Integer

n/a

yes

Seed for the random number generator used during training.

modelName

String

n/a

no

The name of the model to train, must not exist in the Model Catalog.

jobId

String

Generated internally

yes

An ID that can be provided to more easily track the training’s progress.

Table 3. Results
Name Type Description

trainMillis

Integer

Milliseconds used for training.

modelInfo

Map

Information about the training and the winning model.

modelSelectionStats

Map

Statistics about evaluated metrics for all model candidates.

configuration

Map

Configuration used for the train procedure.

The modelInfo can also be retrieved at a later time by using the Model List Procedure. The modelInfo return field has the following algorithm-specific subfields:

Table 4. Model info fields
Name Type Description

bestParameters

Map

The model parameters which performed best on average on validation folds according to the primary metric.

metrics

Map

Map from metric description to evaluated metrics for the winning model over the subsets of the data, see below.

pipeline

Map

The pipeline used to generate and select the node features.

The structure of modelInfo is:

{
    bestParameters: Map,        (1)
    pipeline: Map               (2)
    metrics: {                  (3)
        <METRIC_NAME>: {        (4)
            test: Float,        (5)
            outerTrain: Float,  (6)
            train: {            (7)
                avg: Float,
                max: Float,
                min: Float,
            },
            validation: {       (8)
                avg: Float,
                max: Float,
                min: Float,
                params: Map
            }
        }
    }
}
1 The best scoring model candidate configuration.
2 The pipeline used to generate and select the node features
3 The metrics map contains an entry for each metric description, and the corresponding results for that metric.
4 A metric name specified in the configuration of the procedure, e.g., F1_MACRO or RECALL(class=4).
5 Numeric value for the evaluation of the winning model on the test set.
6 Numeric value for the evaluation of the winning model on the outer train set.
7 The train entry summarizes the metric results over the train set.
8 The validation entry summarizes the metric results over the validation set.

In addition to the data the procedure yields, there’s a fair amount of information about the training that’s being sent to the Neo4j database’s logs as the procedure progresses.

For example, how well each model candidates perform is logged with info log level and thus end up the neo4j.log file of the database.

Some information is only logged with debug log level, and thus end up in the debug.log file of the database. An example of this is training method specific metadata - such as per epoch loss for logistic regression - during model candidate training (in the model selection phase). Please note that this particular data is not yielded by the procedure call.

3. Example

In this section we will show examples of running a Node Regression training pipeline on a concrete graph. The intention is to illustrate what the results look like and to provide a guide in how to make use of the model in a real setting. We will do this on a small graph of a handful of nodes representing houses. In our example we want to predict the price of a house. The example graph looks like this:

node property pipeline graph
The following Cypher statement will create the example graph in the Neo4j database:
CREATE
  (:House {color: 'Gold', sizePerStory: [15.5, 23.6, 33.1], price: 99.99}),
  (:House {color: 'Red', sizePerStory: [15.5, 23.6, 100.0], price: 149.99}),
  (:House {color: 'Blue', sizePerStory: [11.3, 35.1, 22.0], price: 77.77}),
  (:House {color: 'Green', sizePerStory: [23.2, 55.1, 0.0], price: 80.80}),
  (:House {color: 'Gray', sizePerStory: [34.3, 24.0, 0.0],  price: 57.57}),
  (:House {color: 'Black', sizePerStory: [71.66, 55.0, 0.0], price: 140.14}),
  (:House {color: 'White', sizePerStory: [11.1, 111.0, 0.0], price: 122.22}),
  (:House {color: 'Teal', sizePerStory: [80.8, 0.0, 0.0], price: 80.80}),
  (:House {color: 'Beige', sizePerStory: [106.2, 0.0, 0.0], price: 110.11}),
  (:House {color: 'Magenta', sizePerStory: [99.9, 0.0, 0.0], price: 100.00}),
  (:House {color: 'Purple', sizePerStory: [56.5, 0.0, 0.0], price: 60.00}),
  (:UnknownHouse {color: 'Pink', sizePerStory: [23.2, 55.1, 56.1]}),
  (:UnknownHouse {color: 'Tan', sizePerStory: [22.32, 102.0, 0.0]}),
  (:UnknownHouse {color: 'Yellow', sizePerStory: [39.0, 0.0, 0.0]});

With the graph in Neo4j we can now project it into the graph catalog to prepare it for the pipeline execution. We do this using a native projection targeting the House and UnknownHouse labels. We will also project the sizeOfStory property to use as a model feature, and the price property to use as a target feature.

In the examples below we will use named graphs and native projections as the norm. However, Cypher projections can also be used.

The following statement will project a graph using a native projection and store it in the graph catalog under the name 'myGraph'.
CALL gds.graph.project('myGraph', {
    House: { properties: ['sizePerStory', 'price'] },
    UnknownHouse: { properties: 'sizePerStory' }
  },
  '*'
)

3.1. Train

In the following examples we will demonstrate running the Node Regression training pipeline on this graph. We will train a model to predict the price of a house, based on its sizePerStory property. The configuration of the pipeline is the result of running the examples on the previous page:

The following will train a model using a pipeline:
CALL gds.alpha.pipeline.nodeRegression.train('myGraph', {
  pipeline: 'pipe',
  nodeLabels: ['House'],
  modelName: 'nr-pipeline-model',
  targetProperty: 'price',
  randomSeed: 42,
  concurrency: 1,
  metrics: ['MEAN_SQUARED_ERROR']
}) YIELD modelInfo
RETURN
  modelInfo.bestParameters AS winningModel,
  modelInfo.metrics.MEAN_SQUARED_ERROR.train.avg AS avgTrainScore,
  modelInfo.metrics.MEAN_SQUARED_ERROR.outerTrain AS outerTrainScore,
  modelInfo.metrics.MEAN_SQUARED_ERROR.test AS testScore
Table 5. Results
winningModel avgTrainScore outerTrainScore testScore

{maxDepth=2147483647, minLeafSize=1, minSplitSize=2, numberOfDecisionTrees=5, methodName=RandomForest, numberOfSamplesRatio=1.0}

1227.8052857714285

1568.8735900000001

1996.408637333333

Here we can observe that the RandomForest candidate with 5 decision trees performed the best in the training phase. Notice that this is just a toy example on a very small graph. In order to achieve a higher test score, we may need to use better features, a larger graph, or different model configuration.