KNN supports multiple node properties via the nodeProperties key.

Improved normalization of KNN metrics to make them consistent and usable in combination.

Configured with perturbationRate as a value between 0 and 1.

Added perturbation rate to KNN, to reduce the risk of some neighbors not being explored.

Added possibility to exclude pairs of nodes in the K-Nearest Neighbor algorithm that have a similarity below a given threshold defined with an optional configuration parameter similaritењњutoff.

The configuration key initialSampler accepts either UNIFORM or RANDOM_WALK.

Added a random walk sampler for initializing KNN based on the topology of the input graph.

Removed USE_PRE_AGGREGATION feature toggle.The graph name or graph create config always go into the first parameter, the algorithm config always into the second. Changed the syntax of memory estimation.Memory estimation is an exception to this.Only explicit, named graphs are supported. Removed support for anonymous graph syntax from algorithm execution.Inputs for comparison expressions in must resolve to the same type, i.e., long or double.In, defining the same node property for different labels with different neoPropertyKeys is no longer allowed.Removed useBiasFeature parameter in .addLogisticRegression.This also removes the previous default model candidate. Replaced (pipelineName::String, parameterSpace::List of Map) by (pipelineName::String, config::Map.Trained pipelines (which we refer to as models) are still managed in the model catalog. Changed node classification and link prediction training pipelines management from the model catalog to the new pipeline catalog.'scaleProperties' must now be written as 'alpha.scaleProperties'. The procedure resolution for the taskName parameter of. and. changed and now requires the inclusion of the tier, e.g.Removed batchSize parameter for Node Classification pipeline predict modes, because it is not useful.

Additional changes to node classification & link predictions.Removed non-pipeline versions of Link Prediction, including procedures:.Removed non-pipeline versions of Node Classification, including procedures:.Moved node classification pipeline procedures to beta tier:.Moved link prediction pipeline procedures to beta tier:.Renamed YIELD field centrality to score in stream mode.Renamed configuration parameter improved to useWassermanFaust.Removed return item nodes from write and mutate mode.Moved Closeness Centrality to beta tier.Moved delta stepping shortest path to product tier.Removed alpha similarity procedures (similarity metrics added as modes for KNN and Node Similarity).Pearson similarity function no longer accepts Lists of Maps, but computes over Lists of Numbers like the other similarity functions.Moved alpha similarity functions to product tier.nodeWeightProperty for KNN replaced by nodeProperties, which accepts multiple properties.Rename YIELD field startNodeId to sourceNode.Rename configuration parameter startNodeId to sourceNode.Removed configuration parameter relationshipWeightProperty.For a 4.1 or 4.2 compatible release, please see GDS 1.8.6 Breaking changes For a 4.0 compatible release, please see GDS 1.6.5. For a 3.5 compatible release, please see GDS 1.1.7.

Download neo4j graph gallery