Change Password

Input error
Input error
Input error
Submit

Change Nickname

Current Nickname:
Submit
Search
v4.0
    v4.0

    GraphSAGE Train

      AI  

    Related material of the algorithm is as below:

    Command

    algo(graph_sage_train).params(<>)

    Configuration Item Type Default Value Specification Description
    dimension int 64 >0 Dimension of node after the graph embedding, i.e. number of properties
    node_property_names []@<schema>?.<property> / Node property, LTE needed The node property to be used for training
    edge_property_name @<schema>?.<property> (Weight of 1) Numeric edge property, LTE needed Name of the property where the edge weight locates during random walk, schema can be either carried or not; edges without this property will not participate in the degree calculation
    search_depth int 5 >0 Maximum length of steps of random walk
    sample_size []int [25, 10] >0 Maximum number of samples per layer around the start node during random walk
    learning_rate float 0.1 [0, 1] Learning rate of each iteration, the smaller the value, the slower the convergence, larger value leads to more adjustments, but may oscillate due to over-adjustment
    epochs int 10 >0 Number of large cycles
    tolerance tolerance 1e-10 >0 The criterion for judging convergence at the end of each large cycle, it is determined as convergence if less than this value
    aggregator string mean mean or pool mean: Mean aggregator
    pool: Pooling aggregator

    File Writeback

    .write({file: {<>}})

    Parameter Type Default Value Specification Description
    model_name string / / Name of the file path to be written back. Content of the file is the model after being trained.

    Property Writeback

    (Not supported)

    Statistics Writeback

    (Not supported)

    Direct Return

    (Not supported)

    Streaming Return

    (Not supported)

    Real-time Statistics

    (Not supported)

    Please complete the following information to download this book
    *
    公司名称不能为空
    *
    公司邮箱必须填写
    *
    你的名字必须填写
    *
    你的电话必须填写
    *
    你的电话必须填写