Document Vector Applier

This node creates a document vector for each document representing it in the terms space, exactly as the normal document vector node. The difference is that this node takes one data table and one Document Vector model as input:
1. Document Vector model containing names of feature space columns and node settings
2. Table containing the bag-of-words terms

The terms from the first input will be converted into document vectors using the vector from the second input as the reference. Features that appear in first table, but not in the model input will be filtered out, and features that appear in model input, but not in the first table will be added to the output vector and their values will be set to 0.

The values of the feature vectors can be specified as boolean values or as values of a specified column i.e. an tf*idf column. The dimension of the vectors will be the number of distinct terms in the BoW.

Options

Document column

The column containing the documents to use.

Use settings from model

If checked, settings contained in the input model will be used. Note: The settings contained in the input model will node change settings in the node dialog.

Bitvector

If checked a bitvector will be created indicating whether a certain term is contained in a document or not.

Vector value

If Bitvector setting is not checked it is possible to specify the column to use as feature vector values. The column can i.e. contain tf*idf values which are than used as values of the feature vector. Be aware that you have to compute these values before using this node. To do so i.e. the frequency calculation nodes can be used.

As collection cell

If checked all vector entries will be stored in a collection cell consisting of double cells. The cells are ordered, the ordering is specified in the data table spec. If not checked all double cells will be stored in corresponding columns. The advantage of the column representation is that most of the regular algorithms in KNIME can be applied. The disadvantage is (which is on the other hand the advantage of the collection representation) that processing of subsequent nodes will be slowed down, due to the many columns that will be created (dependent on the input data of course).

Feature Column Selection

Selects all feature column names from the model input that should appear in the output document vector.

Views

This node has no views

Workflows

• 01_Emil_the_TeacherBotKNIME Hub
• 01_Initial_Model_TrainingKNIME Hub
• 01_Rest_API_Sentiment_PredictionKNIME Hub
• 01_Rest_API_Sentiment_Prediction_AshKNIME Hub
• 01_Sentiment_Analysis_ML_ApproachKNIME Hub
• Show all 119 workflows

Links

• Input Data Selection Document Classificationby RolandBurger on 2018-12-14

  forum.knime.com/p/51481

Developers