Link Prediction Pipeline - Node properties do not exist

Neo4j Graph Platform Graph Algorithms/Graph Data Science
3

Hi Adam @adam_schill-col !

could you also help me please? I have the same problem an the same error. But my properties are projected in graph.

my graph:

CALL gds.graph.project(
    'myGraph1',
    {
        Project: {
            label: 'Project',
            properties: [
                'projectPlannedRevenue', 
                'projectRevenueRatio', 
                'projectOverallSlippagePercent', 
                'projectCostRatio', 
                'projectTeamSize', 
                'projectDuration', 
                'totalRiskValue', 
                'avgRiskValue', 
                'riskCount', 
                'revenueToCostRatio',
                'embedding'
            ]
        },
        Risk: {
            label: 'Risk',
            properties: [
                'riskValue',  
                'riskDuration',
                'embedding'
            ]
        }
    },
    {
        HAS_RISK: {
            type: 'HAS_RISK',
            orientation: 'UNDIRECTED'
        }
    }
);

my query to train the pipeline:

CALL gds.beta.pipeline.linkPrediction.train('myGraph1', {
  pipeline: 'pipe',
  modelName: 'lp-pipeline-model',
  metrics: ['AUCPR', 'OUT_OF_BAG_ERROR'],
  targetRelationshipType: 'HAS_RISK',
  randomSeed: 12
}) YIELD modelInfo, modelSelectionStats
RETURN
  modelInfo.bestParameters AS winningModel,
  modelInfo.metrics.AUCPR.train.avg AS avgTrainScore,
  modelInfo.metrics.AUCPR.outerTrain AS outerTrainScore,
  modelInfo.metrics.AUCPR.test AS testScore,
  [cand IN modelSelectionStats.modelCandidates | cand.metrics.AUCPR.validation.avg] AS validationScores

image
image1380×320 95.6 KB

(EDIT: also asked in Error by invoking gds.beta.pipeline.linkPrediction.train procedure. Node properties do not exist in the graph. But they do! - #3 by alfeeva.anastasia)