ml_configuration file parameters

Defines the ml_configuration file parameters.

Property Default Value Description
embedding_dim 100 Specifies the smallest dimension that is required to embed an object.
epoch 40 Specifies the number of passes of the entire training data set, the machine learning algorithmmust complete.
hidden_layer_dim 200 Specifies the number of neurons in the hidden layer.
host   Depending upon scenario, can specify the following.GA
Non-SSL MongoDB setup
host=hostname
OR
IP address of the MongoDB
SSL MongoDB setup
host=<host>:<port>/?tls=true&tlsCertificateKeyFile=<mongodb_pem_file_full_path>&tlsCAFile=<mongodb_ca_crt_file_full_path>
Example
XX.XX.XX.XX:27017/?tls=true&tlsCertificateKeyFile=/etc/ssl/mongodb.pem&tlsCAFile=/etc/ssl/mongodb-cert.crt
learning_rate 0.1 Specifies the amount that the weights are updated during training.
name dam_g Specifies the name of the MongoDB database.
number_of_neighbors 2 Specifies the number of the neighbors to consider while constructing metadata (context pairs) for a given input.
password   Specifies the password for the MongoDB database.
port 5000 Depending upon scenario, can specify the port number for any of the following:
  • Controller for the machine learning services
  • MongoDB database
  • Attributes service
  • Categorization service
  • Standardization service
test_split_size 0.30 Specifies the size of the data that must be split as the test data set.
threshold 10 Specifies the threshold that is used while replacing typographical error word from the input.
username mluser Specifies the MongoDB username.