continurous learn or auto machine learn

wang0581

Hi All, I have a trained textmining model, I want to add more exampleset to traning, but I do not want to train everything from zero. Can rapidminer provide continurous learn or auto machine learn? let us traing existing model with more exampleset and classification.  

Thanks!

David_A

Hi @wang0581 ,

you can use the Update Model operator to add new training data to your model. Unfortunately this moslty works with very simply algorithms like Naive Bayes or k-NN and Deep Neural Networks.

Below you find the tutorial process from the Deep Learning extension.

I hope this helps  in our situation.

Best,
David

----------------------------------

<?xml version="1.0" encoding="UTF-8"?><process version="9.5.001">

 <context>

    <input/>
    <output/>
    <macros/>
  </context>
  <operator activated="true" class="process" compatibility="9.4.000" expanded="true" name="Process" origin="GENERATED_TUTORIAL">
    <parameter key="logverbosity" value="init"/>
    <parameter key="random_seed" value="2001"/>
    <parameter key="send_mail" value="never"/>
    <parameter key="notification_email" value=""/>
    <parameter key="process_duration_for_mail" value="30"/>
    <parameter key="encoding" value="SYSTEM"/>
    <process expanded="true">
      <operator activated="true" class="subprocess" compatibility="9.5.001" expanded="true" height="82" name="Subprocess" origin="GENERATED_TUTORIAL" width="90" x="45" y="187">
        <process expanded="true">
          <operator activated="true" class="retrieve" compatibility="9.5.001" expanded="true" height="68" name="Retrieve Abalone" origin="GENERATED_TUTORIAL" width="90" x="112" y="136">
            <parameter key="repository_entry" value="//Samples/Deep Learning/data/Abalone"/>
            <description align="center" color="transparent" colored="false" width="126">Loading data</description>
          </operator>
          <operator activated="true" class="set_role" compatibility="9.5.001" expanded="true" height="82" name="Set Role" origin="GENERATED_TUTORIAL" width="90" x="313" y="136">
            <parameter key="attribute_name" value="Age"/>
            <parameter key="target_role" value="label"/>
            <list key="set_additional_roles"/>
            <description align="center" color="transparent" colored="false" width="126">Choosing 'Age' as the label.</description>
          </operator>
          <operator activated="true" class="nominal_to_numerical" compatibility="9.5.001" expanded="true" height="103" name="Nominal to Numerical" origin="GENERATED_TUTORIAL" width="90" x="514" y="136">
            <parameter key="return_preprocessing_model" value="false"/>
            <parameter key="create_view" value="false"/>
            <parameter key="attribute_filter_type" value="all"/>
            <parameter key="attribute" value=""/>
            <parameter key="attributes" value=""/>
            <parameter key="use_except_expression" value="false"/>
            <parameter key="value_type" value="nominal"/>
            <parameter key="use_value_type_exception" value="false"/>
            <parameter key="except_value_type" value="file_path"/>
            <parameter key="block_type" value="single_value"/>
            <parameter key="use_block_type_exception" value="false"/>
            <parameter key="except_block_type" value="single_value"/>
            <parameter key="invert_selection" value="false"/>
            <parameter key="include_special_attributes" value="false"/>
            <parameter key="coding_type" value="dummy coding"/>
            <parameter key="use_comparison_groups" value="false"/>
            <list key="comparison_groups"/>
            <parameter key="unexpected_value_handling" value="all 0 and warning"/>
            <parameter key="use_underscore_in_name" value="false"/>
          </operator>
          <connect from_op="Retrieve Abalone" from_port="output" to_op="Set Role" to_port="example set input"/>
          <connect from_op="Set Role" from_port="example set output" to_op="Nominal to Numerical" to_port="example set input"/>
          <connect from_op="Nominal to Numerical" from_port="example set output" to_port="out 1"/>
          <portSpacing port="source_in 1" spacing="0"/>
          <portSpacing port="sink_out 1" spacing="0"/>
          <portSpacing port="sink_out 2" spacing="0"/>
          <description align="center" color="purple" colored="true" height="222" resized="true" width="186" x="264" y="48">We want to estimate the 'Age' of sea shells through measuring it.</description>
          <description align="center" color="purple" colored="true" height="222" resized="true" width="186" x="466" y="47">Neural networks only work on numerical data, therefore we need to convert all attributes to numericals.</description>
        </process>
        <description align="center" color="transparent" colored="false" width="126">Load and prepare data</description>
      </operator>
      <operator activated="true" class="split_data" compatibility="9.5.001" expanded="true" height="124" name="Split Data" origin="GENERATED_TUTORIAL" width="90" x="179" y="187">
        <enumeration key="partitions">
          <parameter key="ratio" value="0.2"/>
          <parameter key="ratio" value="0.8"/>
          <parameter key="ratio" value="0.2"/>
        </enumeration>
        <parameter key="sampling_type" value="automatic"/>
        <parameter key="use_local_random_seed" value="false"/>
        <parameter key="local_random_seed" value="1992"/>
        <description align="center" color="transparent" colored="false" width="126">Split data into two training and one test sets.</description>
      </operator>
      <operator activated="true" class="multiply" compatibility="9.5.001" expanded="true" height="103" name="Multiply" origin="GENERATED_TUTORIAL" width="90" x="313" y="340">
        <description align="center" color="transparent" colored="false" width="126">Copy the test set to use the idential data for both model performance checks.</description>
      </operator>
      <operator activated="true" class="deeplearning:dl4j_sequential_neural_network" compatibility="0.9.001" expanded="true" height="103" name="Deep Learning" origin="GENERATED_TUTORIAL" width="90" x="313" y="85">
        <parameter key="loss_function" value="Mean Squared Error (Linear Regression)"/>
        <parameter key="epochs" value="50"/>
        <parameter key="use_miniBatch" value="true"/>
        <parameter key="batch_size" value="4"/>
        <parameter key="updater" value="Adam"/>
        <parameter key="learning_rate" value="0.005"/>
        <parameter key="momentum" value="0.9"/>
        <parameter key="rho" value="0.95"/>
        <parameter key="epsilon" value="1.0E-6"/>
        <parameter key="beta1" value="0.9"/>
        <parameter key="beta2" value="0.999"/>
        <parameter key="RMSdecay" value="0.95"/>
        <parameter key="weight_initialization" value="Xavier"/>
        <parameter key="bias_initialization" value="0.0"/>
        <parameter key="use_regularization" value="false"/>
        <parameter key="l1_strength" value="0.1"/>
        <parameter key="l2_strength" value="0.1"/>
        <parameter key="optimization_method" value="Stochastic Gradient Descent"/>
        <parameter key="backpropagation" value="Standard"/>
        <parameter key="backpropagation_length" value="50"/>
        <parameter key="infer_input_shape" value="true"/>
        <parameter key="network_type" value="Simple Neural Network"/>
        <parameter key="log_each_epoch" value="true"/>
        <parameter key="epochs_per_log" value="10"/>
        <parameter key="use_local_random_seed" value="false"/>
        <parameter key="local_random_seed" value="1992"/>
        <process expanded="true">
          <operator activated="true" class="deeplearning:dl4j_dense_layer" compatibility="0.9.001" expanded="true" height="68" name="Add Fully-Connected Layer" origin="GENERATED_TUTORIAL" width="90" x="112" y="85">
            <parameter key="number_of_neurons" value="8"/>
            <parameter key="activation_function" value="ReLU (Rectified Linear Unit)"/>
            <parameter key="use_dropout" value="false"/>
            <parameter key="dropout_rate" value="0.25"/>
            <parameter key="overwrite_networks_weight_initialization" value="false"/>
            <parameter key="weight_initialization" value="Normal"/>
            <parameter key="overwrite_networks_bias_initialization" value="false"/>
            <parameter key="bias_initialization" value="0.0"/>
          </operator>
          <operator activated="true" class="deeplearning:dl4j_dense_layer" compatibility="0.9.001" expanded="true" height="68" name="Add Fully-Connected Layer (2)" origin="GENERATED_TUTORIAL" width="90" x="313" y="85">
            <parameter key="number_of_neurons" value="1"/>
            <parameter key="activation_function" value="None (identity)"/>
            <parameter key="use_dropout" value="false"/>
            <parameter key="dropout_rate" value="0.25"/>
            <parameter key="overwrite_networks_weight_initialization" value="false"/>
            <parameter key="weight_initialization" value="Normal"/>
            <parameter key="overwrite_networks_bias_initialization" value="false"/>
            <parameter key="bias_initialization" value="0.0"/>
          </operator>
          <connect from_port="layerArchitecture" to_op="Add Fully-Connected Layer" to_port="layerArchitecture"/>
          <connect from_op="Add Fully-Connected Layer" from_port="layerArchitecture" to_op="Add Fully-Connected Layer (2)" to_port="layerArchitecture"/>
          <connect from_op="Add Fully-Connected Layer (2)" from_port="layerArchitecture" to_port="layerArchitecture"/>
          <portSpacing port="source_layerArchitecture" spacing="0"/>
          <portSpacing port="sink_layerArchitecture" spacing="0"/>
        </process>
      </operator>
      <operator activated="true" class="apply_model" compatibility="9.5.001" expanded="true" height="82" name="Apply Model" origin="GENERATED_TUTORIAL" width="90" x="447" y="85">
        <list key="application_parameters"/>
        <parameter key="create_view" value="false"/>
      </operator>
      <operator activated="true" class="performance_regression" compatibility="9.5.001" expanded="true" height="82" name="Performance" origin="GENERATED_TUTORIAL" width="90" x="648" y="85">
        <parameter key="main_criterion" value="first"/>
        <parameter key="root_mean_squared_error" value="false"/>
        <parameter key="absolute_error" value="false"/>
        <parameter key="relative_error" value="true"/>
        <parameter key="relative_error_lenient" value="false"/>
        <parameter key="relative_error_strict" value="false"/>
        <parameter key="normalized_absolute_error" value="false"/>
        <parameter key="root_relative_squared_error" value="false"/>
        <parameter key="squared_error" value="false"/>
        <parameter key="correlation" value="false"/>
        <parameter key="squared_correlation" value="false"/>
        <parameter key="prediction_average" value="false"/>
        <parameter key="spearman_rho" value="false"/>
        <parameter key="kendall_tau" value="false"/>
        <parameter key="skip_undefined_labels" value="true"/>
        <parameter key="use_example_weights" value="true"/>
        <description align="center" color="transparent" colored="false" width="126">Calculate model performance</description>
      </operator>
      <operator activated="true" class="update_model" compatibility="9.5.001" expanded="true" height="82" name="Update Model" origin="GENERATED_TUTORIAL" width="90" x="648" y="289"/>
      <operator activated="true" class="apply_model" compatibility="9.5.001" expanded="true" height="82" name="Apply Model (2)" origin="GENERATED_TUTORIAL" width="90" x="782" y="340">
        <list key="application_parameters"/>
        <parameter key="create_view" value="false"/>
      </operator>
      <operator activated="true" class="performance_regression" compatibility="9.5.001" expanded="true" height="82" name="Performance after Update" origin="GENERATED_TUTORIAL" width="90" x="916" y="340">
        <parameter key="main_criterion" value="first"/>
        <parameter key="root_mean_squared_error" value="false"/>
        <parameter key="absolute_error" value="false"/>
        <parameter key="relative_error" value="true"/>
        <parameter key="relative_error_lenient" value="false"/>
        <parameter key="relative_error_strict" value="false"/>
        <parameter key="normalized_absolute_error" value="false"/>
        <parameter key="root_relative_squared_error" value="false"/>
        <parameter key="squared_error" value="false"/>
        <parameter key="correlation" value="false"/>
        <parameter key="squared_correlation" value="false"/>
        <parameter key="prediction_average" value="false"/>
        <parameter key="spearman_rho" value="false"/>
        <parameter key="kendall_tau" value="false"/>
        <parameter key="skip_undefined_labels" value="true"/>
        <parameter key="use_example_weights" value="true"/>
        <description align="center" color="transparent" colored="false" width="126">Calculate model performance</description>
      </operator>
      <connect from_op="Subprocess" from_port="out 1" to_op="Split Data" to_port="example set"/>
      <connect from_op="Split Data" from_port="partition 1" to_op="Deep Learning" to_port="training set"/>
      <connect from_op="Split Data" from_port="partition 2" to_op="Update Model" to_port="example set"/>
      <connect from_op="Split Data" from_port="partition 3" to_op="Multiply" to_port="input"/>
      <connect from_op="Multiply" from_port="output 1" to_op="Apply Model" to_port="unlabelled data"/>
      <connect from_op="Multiply" from_port="output 2" to_op="Apply Model (2)" to_port="unlabelled data"/>
      <connect from_op="Deep Learning" from_port="model" to_op="Apply Model" to_port="model"/>
      <connect from_op="Apply Model" from_port="labelled data" to_op="Performance" to_port="labelled data"/>
      <connect from_op="Apply Model" from_port="model" to_op="Update Model" to_port="model"/>
      <connect from_op="Performance" from_port="performance" to_port="result 1"/>
      <connect from_op="Update Model" from_port="model" to_op="Apply Model (2)" to_port="model"/>
      <connect from_op="Apply Model (2)" from_port="labelled data" to_op="Performance after Update" to_port="labelled data"/>
      <connect from_op="Performance after Update" from_port="performance" to_port="result 2"/>
      <portSpacing port="source_input 1" spacing="0"/>
      <portSpacing port="sink_result 1" spacing="0"/>
      <portSpacing port="sink_result 2" spacing="0"/>
      <portSpacing port="sink_result 3" spacing="0"/>
      <description align="center" color="yellow" colored="false" height="54" resized="true" width="220" x="34" y="40">Update a previously trained model with new data.</description>
      <description align="center" color="green" colored="true" height="196" resized="true" width="474" x="282" y="40">Initial training &amp;amp; model evaluation</description>
      <description align="center" color="blue" colored="true" height="232" resized="true" width="415" x="616" y="253">Perform one more iteration step and evaluate the model again.</description>
    </process>
  </operator>
</process>