Visualize HashGNN and node2vec node embeddings

JohT · JohT · commit a89db13caaf1 · 2024-05-08T21:51:24.000+02:00
diff --git a/cypher/Node_Embeddings/Node_Embeddings_1a_Fast_Random_Projection_Estimate.cypher b/cypher/Node_Embeddings/Node_Embeddings_1a_Fast_Random_Projection_Estimate.cypher
@@ -3,6 +3,7 @@
 CALL gds.fastRP.stream.estimate(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )
diff --git a/cypher/Node_Embeddings/Node_Embeddings_1b_Fast_Random_Projection_Statistics.cypher b/cypher/Node_Embeddings/Node_Embeddings_1b_Fast_Random_Projection_Statistics.cypher
@@ -3,6 +3,7 @@
 CALL gds.fastRP.stats(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )
diff --git a/cypher/Node_Embeddings/Node_Embeddings_1c_Fast_Random_Projection_Mutate.cypher b/cypher/Node_Embeddings/Node_Embeddings_1c_Fast_Random_Projection_Mutate.cypher
@@ -3,6 +3,7 @@
 CALL gds.fastRP.mutate(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,mutateProperty: $dependencies_projection_write_property
   }
diff --git a/cypher/Node_Embeddings/Node_Embeddings_1d_Fast_Random_Projection_Stream.cypher b/cypher/Node_Embeddings/Node_Embeddings_1d_Fast_Random_Projection_Stream.cypher
@@ -3,6 +3,7 @@
 CALL gds.fastRP.stream(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )
diff --git a/cypher/Node_Embeddings/Node_Embeddings_1e_Fast_Random_Projection_Write.cypher b/cypher/Node_Embeddings/Node_Embeddings_1e_Fast_Random_Projection_Write.cypher
@@ -3,6 +3,7 @@
 CALL gds.fastRP.write(
  $dependencies_projection + '-cleaned', {
       embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,writeProperty: $dependencies_projection_write_property
   }
diff --git a/cypher/Node_Embeddings/Node_Embeddings_2a_Hash_GNN_Estimate.cypher b/cypher/Node_Embeddings/Node_Embeddings_2a_Hash_GNN_Estimate.cypher
@@ -6,9 +6,11 @@ CALL gds.beta.hashgnn.stream.estimate(
      ,iterations: 3
      ,generateFeatures: {
          dimension: toInteger($dependencies_projection_embedding_dimension) * 4
-        ,densityLevel: 1
+        ,densityLevel: 3
      }
      ,outputDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,neighborInfluence: 0.9
+     ,randomSeed: 30
   }
 )
  YIELD requiredMemory, nodeCount, relationshipCount, bytesMin, bytesMax, heapPercentageMin, heapPercentageMax, treeView
diff --git a/cypher/Node_Embeddings/Node_Embeddings_2c_Hash_GNN_Mutate.cypher b/cypher/Node_Embeddings/Node_Embeddings_2c_Hash_GNN_Mutate.cypher
@@ -6,9 +6,11 @@ CALL gds.beta.hashgnn.mutate(
      ,iterations: 3
      ,generateFeatures: {
          dimension: toInteger($dependencies_projection_embedding_dimension) * 4
-        ,densityLevel: 1
+        ,densityLevel: 3
      }
      ,outputDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,neighborInfluence: 0.9
+     ,randomSeed: 30
      ,mutateProperty: $dependencies_projection_write_property
   }
 )
diff --git a/cypher/Node_Embeddings/Node_Embeddings_2d_Hash_GNN_Stream.cypher b/cypher/Node_Embeddings/Node_Embeddings_2d_Hash_GNN_Stream.cypher
@@ -6,9 +6,11 @@ CALL gds.beta.hashgnn.stream(
      ,iterations: 3
      ,generateFeatures: {
          dimension: toInteger($dependencies_projection_embedding_dimension) * 4
-        ,densityLevel: 1
+        ,densityLevel: 3
      }
      ,outputDimension: toInteger($dependencies_projection_embedding_dimension)
+     ,neighborInfluence: 0.9
+     ,randomSeed: 30
   }
 )
 YIELD nodeId, embedding
diff --git a/cypher/Node_Embeddings/Node_Embeddings_3a_Node2Vec_Estimate.cypher b/cypher/Node_Embeddings/Node_Embeddings_3a_Node2Vec_Estimate.cypher
@@ -4,6 +4,7 @@ CALL gds.node2vec.write.estimate(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,writeProperty: $dependencies_projection_write_property
   }
diff --git a/cypher/Node_Embeddings/Node_Embeddings_3c_Node2Vec_Mutate.cypher b/cypher/Node_Embeddings/Node_Embeddings_3c_Node2Vec_Mutate.cypher
@@ -4,6 +4,7 @@ CALL gds.node2vec.mutate(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,mutateProperty: $dependencies_projection_write_property
   }
diff --git a/cypher/Node_Embeddings/Node_Embeddings_3d_Node2Vec_Stream.cypher b/cypher/Node_Embeddings/Node_Embeddings_3d_Node2Vec_Stream.cypher
@@ -4,6 +4,7 @@ CALL gds.node2vec.stream(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
   }
 )
diff --git a/cypher/Node_Embeddings/Node_Embeddings_3e_Node2Vec_Write.cypher b/cypher/Node_Embeddings/Node_Embeddings_3e_Node2Vec_Write.cypher
@@ -4,6 +4,7 @@ CALL gds.node2vec.write(
  $dependencies_projection + '-cleaned', {
      ,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)
      ,iterations: 3
+     ,randomSeed: 30
      ,relationshipWeightProperty: $dependencies_projection_weight_property
      ,writeProperty: $dependencies_projection_write_property
   }
diff --git a/jupyter/NodeEmbeddings.ipynb b/jupyter/NodeEmbeddings.ipynb

Original file line number	Diff line number	Diff line change
`@@ -3,6 +3,7 @@`
`3`	`3`	`CALL gds.fastRP.stream.estimate(`
`4`	`4`	`$dependencies_projection + '-cleaned', {`
`5`	`5`	`embeddingDimension: toInteger($dependencies_projection_embedding_dimension)`
	`6`	`+ ,randomSeed: 30`
`6`	`7`	`,relationshipWeightProperty: $dependencies_projection_weight_property`
`7`	`8`	`}`
`8`	`9`	`)`
Original file line number	Diff line number	Diff line change
`@@ -3,6 +3,7 @@`
`3`	`3`	`CALL gds.fastRP.stats(`
`4`	`4`	`$dependencies_projection + '-cleaned', {`
`5`	`5`	`embeddingDimension: toInteger($dependencies_projection_embedding_dimension)`
	`6`	`+ ,randomSeed: 30`
`6`	`7`	`,relationshipWeightProperty: $dependencies_projection_weight_property`
`7`	`8`	`}`
`8`	`9`	`)`
Original file line number	Diff line number	Diff line change
`@@ -3,6 +3,7 @@`
`3`	`3`	`CALL gds.fastRP.mutate(`
`4`	`4`	`$dependencies_projection + '-cleaned', {`
`5`	`5`	`embeddingDimension: toInteger($dependencies_projection_embedding_dimension)`
	`6`	`+ ,randomSeed: 30`
`6`	`7`	`,relationshipWeightProperty: $dependencies_projection_weight_property`
`7`	`8`	`,mutateProperty: $dependencies_projection_write_property`
`8`	`9`	`}`
Original file line number	Diff line number	Diff line change
`@@ -3,6 +3,7 @@`
`3`	`3`	`CALL gds.fastRP.stream(`
`4`	`4`	`$dependencies_projection + '-cleaned', {`
`5`	`5`	`embeddingDimension: toInteger($dependencies_projection_embedding_dimension)`
	`6`	`+ ,randomSeed: 30`
`6`	`7`	`,relationshipWeightProperty: $dependencies_projection_weight_property`
`7`	`8`	`}`
`8`	`9`	`)`
Original file line number	Diff line number	Diff line change
`@@ -3,6 +3,7 @@`
`3`	`3`	`CALL gds.fastRP.write(`
`4`	`4`	`$dependencies_projection + '-cleaned', {`
`5`	`5`	`embeddingDimension: toInteger($dependencies_projection_embedding_dimension)`
	`6`	`+ ,randomSeed: 30`
`6`	`7`	`,relationshipWeightProperty: $dependencies_projection_weight_property`
`7`	`8`	`,writeProperty: $dependencies_projection_write_property`
`8`	`9`	`}`
Original file line number	Diff line number	Diff line change
`@@ -6,9 +6,11 @@ CALL gds.beta.hashgnn.stream.estimate(`
`6`	`6`	`,iterations: 3`
`7`	`7`	`,generateFeatures: {`
`8`	`8`	`dimension: toInteger($dependencies_projection_embedding_dimension) * 4`
`9`		`- ,densityLevel: 1`
	`9`	`+ ,densityLevel: 3`
`10`	`10`	`}`
`11`	`11`	`,outputDimension: toInteger($dependencies_projection_embedding_dimension)`
	`12`	`+ ,neighborInfluence: 0.9`
	`13`	`+ ,randomSeed: 30`
`12`	`14`	`}`
`13`	`15`	`)`
`14`	`16`	`YIELD requiredMemory, nodeCount, relationshipCount, bytesMin, bytesMax, heapPercentageMin, heapPercentageMax, treeView`
Original file line number	Diff line number	Diff line change
`@@ -4,6 +4,7 @@ CALL gds.node2vec.write.estimate(`
`4`	`4`	`$dependencies_projection + '-cleaned', {`
`5`	`5`	`,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)`
`6`	`6`	`,iterations: 3`
	`7`	`+ ,randomSeed: 30`
`7`	`8`	`,relationshipWeightProperty: $dependencies_projection_weight_property`
`8`	`9`	`,writeProperty: $dependencies_projection_write_property`
`9`	`10`	`}`
Original file line number	Diff line number	Diff line change
`@@ -4,6 +4,7 @@ CALL gds.node2vec.mutate(`
`4`	`4`	`$dependencies_projection + '-cleaned', {`
`5`	`5`	`,embeddingDimension: toInteger($dependencies_projection_embedding_dimension)`
`6`	`6`	`,iterations: 3`
	`7`	`+ ,randomSeed: 30`
`7`	`8`	`,relationshipWeightProperty: $dependencies_projection_weight_property`
`8`	`9`	`,mutateProperty: $dependencies_projection_write_property`
`9`	`10`	`}`