Classify New Queries

You can predict the categories most likely to satisfy a new query using this workflow:

  1. Use the Build Training Data job to join your signals data with your catalog data and produce training data in the form of query/class pairs.

  2. Use the Classification job to train a classification model using the output collection of the Build Training Data job as the training collection.

  3. Specify the classification model’s name in the Machine Learning stage of your query pipeline.

Query-time classification workflow

Query-time classification workflow

See the detailed steps below.

To predict the categories of new queries

  1. Navigate to Collections > Jobs > Add+ > Build Training Data to create a new Build Training Data job.

  2. Configure the job as follows:

    1. In the Catalog Path field, enter the collection name or cloud storage path where your main content is stored.

    2. In the Catalog Format field, enter solr if you are analyzing a Solr collection, or another format if your content is in the cloud.

    3. In the Signals Path field, enter the collection name or cloud storage path where your signals data is stored.

    4. In the Output Path field, enter the collection name or cloud storage path where you want to store the training data.

    5. In the Category Field in Catalog field, enter the field name for the category data in your main content.

    6. In the Item ID Field in Catalog field, enter the field name for the item IDs in your main content.

    7. Check that the values of Item ID Field in Signals and Count Field in Signals match the field names in your signals data.

  3. Save the job.

  4. Click Run > Start to run the job.

  5. Navigate to Collections > Jobs > Add+ > Classification to create a new Classification job.

  6. Configure the job as follows:

    1. In the Model Deployment Name field, enter an ID for the new classification model.

    2. In the Training Data Path field, enter the collection name or cloud storage path from the Build Training Data job’s Output Path field.

    3. In the Training Data Format field, leave the default solr value if the Training Data Path is a collection or if you used the default format in your Build Training Data job configuration.

      If you configured the Build Training Data job to output a different format, enter it here.

    4. In the Training collection content field, enter query_s, the default content field name in the Build Training Data job’s output.

    5. In the Training collection class field, enter category_s, the default category field name in the Build Training Data job’s output.

      Tip
      For additional configuration details, see Best practices below.
  7. Save the job.

  8. Verify that the Build Training Data job has finished successfully.

  9. Click Run > Start to run the job.

  10. Navigate to Indexing > Query Workbench > Load and select your query pipeline.

  11. Configure the query pipeline as follows:

    1. Add a new Machine Learning stage.

    2. In the Model ID field, enter the name from the Classification job’s Model Deployment Name field.

    3. In the Model input transformation script field, enter the following:

      var modelInput = new java.util.HashMap()
      modelInput.put("text", request.getFirstParam("q"))
      modelInput
    4. In the Model output transformation script field, enter the following:

      // In case if top_k_predictions are needed
      // To put into response documents (can be done only after Solr Query stage)
      var jsonOutput = JSON.parse(modelOutput.get("_rawJsonResponse"))
      var parsedOutput = {};
      for (var i=0; i<jsonOutput["names"].length;i++){
        parsedOutput[jsonOutput["names"][i]] = jsonOutput["ndarray"][i]
      }
      
      var docs = response.get().getInnerResponse().getDocuments();
      var ndocs = new java.util.ArrayList();
      for (var i=0; i<docs.length;i++){
        var doc = docs[i];
        doc.putField("top_1_class", parsedOutput["top_1_class"][0])
        doc.putField("top_1_score", parsedOutput["top_1_score"][0])
        if ("top_k_classes" in parsedOutput) {
          doc.putField("top_k_classes", new java.util.ArrayList(parsedOutput["top_k_classes"][0]))
          doc.putField("top_k_scores", new java.util.ArrayList(parsedOutput["top_k_scores"][0]))
        }
        ndocs.add(doc);
      }
      response.get().getInnerResponse().updateDocuments(ndocs);
    5. Click Apply.

  1. Save the query pipeline.

Custom output transformation script examples

// To put into request
request.putSingleParam("class", modelOutput.get("top_1_class")[0])
request.putSingleParam("score", modelOutput.get("top_1_score")[0])

// Or for example to apply Filter Query
request.putSingleParam("fq", "class:" + modelOutput.get("top_1_class")[0])
// To put into query context
context.put("class", modelOutput.get("top_1_class")[0])
context.put("score", modelOutput.get("top_1_score")[0])
// To put into response documents (can be done only after Solr Query stage)
var docs = response.get().getInnerResponse().getDocuments();
var ndocs = new java.util.ArrayList();

for (var i=0; i<docs.length;i++){
  var doc = docs[i];
  doc.putField("query_class", modelOutput.get("top_1_class")[0])
  doc.putField("query_score", modelOutput.get("top_1_score")[0])
  ndocs.add(doc);
}

response.get().getInnerResponse().updateDocuments(ndocs);

Best practices for configuring the Classification job

These tips describe how to tune the options under Vectorization Parameters for best results with different use cases.

Query intent / short texts

If you want to train a model to predict query intents or to do short text classification, then enable Use Characters.

Another vectorization parameter that can improve model quality is Max Ngram size, with reasonable defaults between 3 and 5.

The more character ngrams are used the bigger the vocabulary, so it is worthwhile to tune the Maximum Vocab Size parameter that controls how many unique tokens will be used. Lower values will make training faster and will prevent overfitting but might provide lower quality too. It’s important to find a good balance.

Activating the advanced Sublinear TF option usually helps if characters are used.

Documents / long texts

If you want to train a model to predict classes for documents or long texts like one or more paragraphs, then uncheck Use Characters.

The reasonable values for word-based Max Ngram size are 2–3. Be sure to tune Maximum Vocab Size parameter too. Usually it’s better to leave the advanced Sublinear TF option deactivated.

Performance tuning

If the text is very long and Use Characters is checked, the job may take a lot of memory and possibly fail if the amount of memory requested by the job is not available. This may result in pods being evicted or failing with OOM errors. If you see this happening, try the following:

  • Uncheck Use Characters.

  • Reduce the vocabulary size and ngram range of the documents.

  • Allocate more memory to the pod.

Algorithm-specific

If you are going to train a model via LogisticRegression algorithm, dimensionality reduction usually doesn’t help so it makes sense to leave Reduce Dimensionality unchecked. But scaling seems to improve results, so it’s suggested to activate Scale Features.

For models trained by StarSpace algorithm it’s vice-versa. Dimensionality reduction usually helps to get better results as well as much faster model training. But scaling usually doesn’t help or might make results a little bit worse.