Icon

Customer Opinion Analysis with Redfield BERT and Spacy nodes

Document tokenization and spaCy tagging for document entities: Spacy nodes have been used to test automatic NLP techniques and models (dictionaries) over Spanish Language as lexical and semanticstagging, morphologiging entities and others. CUSTOMER FEEDBACK OPINIONS ANALYSIS WITH BERT AND SPACY NODES BY REDFIELD - (SPANISH LANGUAGE DATA)Text analysis has become a mayor field in Digital Analytics and Digital Marketing, as unstructured text content is almost everywhere.This workflow shows an example of applying NLP techniques to a customer reviews from an online course, a starting point to reduce "human time" in these high-time consumption tasks.All information about BERT and Spacy nodes by Redfield here: https://hub.knime.com/redfield/extensions/se.redfield.bert.feature/latest DATA ENTRY: Opinions from "FashionDesing Online Course"(example - randomized) Training BERT Model without text data pre - processing: BERT Model nodes by Redfield can manage text variables without any previous data transformation. This feature simplify greatly the time needed to start deploying NPL models for text modelling. DOCUMENT - TYPETRANSFORMMCM - OPINIONES(RESUMIDO)POINT OF SPEECH TAGGINGNode 901Node 902MORPH TAGGERCONTENT / TEXT"NORMALIZATION"Node 905NAMES TAGGERNode 910Node 927PREDICTION:TRESHOLD ADJUST:>0.33 for "si" CATEGORYNode 934Node 938CLASS & SENTENCE COLUMN SELECTION - BERT MODEL ADJUSTNEEDS A LOCAL FOLDER TO UPLOAD BERT MODELSELECTEDNode 942subsets for train / test:80% / 20%(tricky part to multiplyregs to train BERT model) Spacy Tokenizer File Reader Spacy POS Tagger Document Viewer Document Viewer Spacy Morphologizer Spacy Lemmatizer Document Viewer Spacy NER Document Viewer Scorer Rule Engine Column Filter Shuffle BERT ClassificationLearner BERT Model Selector BERT Predictor Partitioning Bootstrap Sampling Document tokenization and spaCy tagging for document entities: Spacy nodes have been used to test automatic NLP techniques and models (dictionaries) over Spanish Language as lexical and semanticstagging, morphologiging entities and others. CUSTOMER FEEDBACK OPINIONS ANALYSIS WITH BERT AND SPACY NODES BY REDFIELD - (SPANISH LANGUAGE DATA)Text analysis has become a mayor field in Digital Analytics and Digital Marketing, as unstructured text content is almost everywhere.This workflow shows an example of applying NLP techniques to a customer reviews from an online course, a starting point to reduce "human time" in these high-time consumption tasks.All information about BERT and Spacy nodes by Redfield here: https://hub.knime.com/redfield/extensions/se.redfield.bert.feature/latest DATA ENTRY: Opinions from "FashionDesing Online Course"(example - randomized) Training BERT Model without text data pre - processing: BERT Model nodes by Redfield can manage text variables without any previous data transformation. This feature simplify greatly the time needed to start deploying NPL models for text modelling. DOCUMENT - TYPETRANSFORMMCM - OPINIONES(RESUMIDO)POINT OF SPEECH TAGGINGNode 901Node 902MORPH TAGGERCONTENT / TEXT"NORMALIZATION"Node 905NAMES TAGGERNode 910Node 927PREDICTION:TRESHOLD ADJUST:>0.33 for "si" CATEGORYNode 934Node 938CLASS & SENTENCE COLUMN SELECTION - BERT MODEL ADJUSTNEEDS A LOCAL FOLDER TO UPLOAD BERT MODELSELECTEDNode 942subsets for train / test:80% / 20%(tricky part to multiplyregs to train BERT model) Spacy Tokenizer File Reader Spacy POS Tagger Document Viewer Document Viewer Spacy Morphologizer Spacy Lemmatizer Document Viewer Spacy NER Document Viewer Scorer Rule Engine Column Filter Shuffle BERT ClassificationLearner BERT Model Selector BERT Predictor Partitioning Bootstrap Sampling

Nodes

Extensions

Links