17.4.4 Supported Property Types for Unsupervised GraphWise Model

The model supports two types of properties for both vertices and edges:

• continuous properties (boolean, double, float, integer, long)
• categorical properties (string)

For categorical properties, two categorical configurations are possible:

• One-hot-encoding: Each category is mapped to a vector, that is concatenated to other features (default)
• Embedding table: Each category is mapped to an embedding that is concatenated to other features and is trained along with the model

One-hot-encoding converts each category into an independent vector. This is useful if you want each category to be interpreted as an equally independent group. For instance, if there are categories ranging from A to E, where each alphabet has no specific meaning, then one-hot-encoding can be a good fit.

Embedding table is recommended if the semantics of the properties matter, and you want certain categories to be closer to each other than the others. For example, assume there is a day property with values ranging from Monday to Sunday. If you wish to preserve the idea that Tuesday is closer to Wednesday than Saturday, then by choosing the embedding table configuration, you can let the vectors that represent the categories to be learned during training, so that the vector that is mapped to Tuesday becomes close to that of Wednesday.

One advantage that the embedding table approach has over one-hot-encoding is that you can learn more suitable vectors to represent each category. However, this also means that a good amount of data is required to train the embedding table properly. The one-hot-encoding approach might be better for use-cases with limited training data.

When using the embedding table, users are allowed to set the out-of-vocabulary probability. With the given probability, the embedding will be set to the out-of-vocabulary embedding randomly during training, in order to make the model more robust to unseen categories during inference.

• JShell
• Java
• Python

opg4j> import oracle.pgx.config.mllib.inputconfig.CategoricalPropertyConfig;
opg4j> var prop1config = analyst.categoricalPropertyConfigBuilder("vertex_str_feature_1").
    oneHotEncoding().
    setMaxVocabularySize(100).
    build()
opg4j> var prop2config = analyst.categoricalPropertyConfigBuilder("vertex_str_feature_2").
    embeddingTable().
    setShared(false). // set whether to share the vocabulary or not when several vertex types have a property with the same name
    setEmbeddingDimension(32).
    setOutOfVocabularyProbability(0.001). // probability to set the word embedding to the out-of-vocabulary embedding
    build()
opg4j> var model = analyst.unsupervisedGraphWiseModelBuilder().
    setVertexInputPropertyNames(
        "vertex_int_feature_1", // continuous feature
        "vertex_str_feature_1", // string feature using one-hot-encoding
        "vertex_str_feature_2", // string feature using embedding table
        "vertex_str_feature_3" // string feature using one-hot-encoding (default)
    ).
    setVertexInputPropertyConfigs(prop1config, prop2config).
    build()

import oracle.pgx.config.mllib.inputconfig.CategoricalPropertyConfig;
import oracle.pgx.config.mllib.inputconfig.InputPropertyConfig;

InputPropertyConfig prop1config = analyst.categoricalPropertyConfigBuilder("vertex_str_feature_1")
    .oneHotEncoding()
    .setMaxVocabularySize(100)
    .build();
InputPropertyConfig prop2config = analyst.categoricalPropertyConfigBuilder("vertex_str_feature_2")
    .embeddingTable()
    .setShared(false) // set whether to share the vocabulary or not when several vertex types have a property with the same name
    .setEmbeddingDimension(32)
    .setOutOfVocabularyProbability(0.001) // probability to set the word embedding to the out-of-vocabulary embedding
    .build();
SupervisedGraphWiseModelBuilder model = analyst.unsupervisedGraphWiseModelBuilder()
    .setVertexInputPropertyNames(
        "vertex_int_feature_1", // continuous feature
        "vertex_str_feature_1", // string feature using one-hot-encoding
        "vertex_str_feature_2", // string feature using embedding table
        "vertex_str_feature_3" // string feature using one-hot-encoding (default)
    )
    .setVertexInputPropertyConfigs(prop1config, prop2config)
    .build();

vertex_input_property_configs = [
    analyst.one_hot_encoding_categorical_property_config(
        property_name="vertex_str_feature_1",
        max_vocabulary_size=100,
    ),
    analyst.learned_embedding_categorical_property_config(
        property_name="vertex_str_feature_2",
        embedding_dim=4,
        shared=False, // set whether to share the vocabulary or not when several  types have a property with the same name
        oov_probability=0.001 // probability to set the word embedding to the out-of-vocabulary embedding
    )
]

model_params = dict(
    vertex_input_property_names=[
        "vertex_int_feature_1", // continuous feature
        "vertex_str_feature_1", // string feature using one-hot-encoding
        "vertex_str_feature_2", // string feature using embedding table
        "vertex_str_feature_3", // string feature using one-hot-encoding (default)
    ],
    vertex_input_property_configs=vertex_input_property_configs
)

model = analyst.supervised_graphwise_builder(**model_params)