Deep Learning
Classify texts with an OpenNLP naive Bayes model: build a training file from labelled lines, train and load a model, then predict the category of a text. Typical uses: route emails or tickets by subject.
dl n_bayesian create_train_filedl n_bayesian showdl n_bayesian existdl n_bayesian create_modeldl n_bayesian loaddl n_bayesian predictdl n_bayesian delete

dl n_bayesian create_train_file


Description

    To get a training file.

admin
file create "demo/train.txt" "transfer VIR RECU 854526,VIR RECU 8545269324 DE: OCTO TECHNOLOGY MOTIF: VIREMENT SALAIRES JUIN 11 withdrawal CARTE X2052 RETRAI,CARTE X2052 RETRAIT DAB 22/06 21H37 CIC PARIS SAINT ROCH 10859A00 payment CARTE X2052 21/06 ,CARTE X2052 21/06 PHIE LA BOETIE fees OPTION TRANQUILLIT,OPTION TRANQUILLITE fees COTISATION JAZZ,COTISATION JAZZ atmfees FRAIS PAIEMENT HOR,FRAIS PAIEMENT HORS ZONE EURO 1 PAIEMENT A 1.00 EUR NT 38.06 EUR A 2.70 atmfees FRAIS PAIEMENT HOR,FRAIS PAIEMENT HORS ZONE EURO 1 PAIEMENT A 1.00 EUR NT 3.33 EUR A";
mentdb

dl n_bayesian show


Description

    To show all Naive Bayesian networks.

admin
dl n_bayesian show
mentdb
[]
Used in use cases (1 recipe)

    Deep learning engineer › Categorise bank statement lines
Often used with

    dl n_bayesian create_model  ·  dl n_bayesian delete  ·  dl n_bayesian exist  ·  dl n_bayesian load  ·  dl n_bayesian predict  ·  file create  ·  file delete

dl n_bayesian exist <key>


Description

    To check if a Naive Bayesian network already exists.

Parameters

    key:   The network key - string - required
admin
dl n_bayesian exist "bayesian1";
mentdb
1
Used in use cases (3 recipes)

    Deep learning engineer › Categorise bank statement lines
    Deep learning engineer › Sentiment triage of customer reviews
    Deep learning engineer › Confusion matrix of an intent classifier
Often used with

    dl n_bayesian create_model  ·  dl n_bayesian delete  ·  dl n_bayesian load  ·  dl n_bayesian predict  ·  file create  ·  file delete  ·  dl n_bayesian show  ·  json doc

dl n_bayesian create_model <lang> <train_file_path> <iterations_param> <model_file_path_to_save>


Description

    Train a naive Bayes text classifier (OpenNLP document categorizer) from a training file and save it in a model file.
The training file has one example per line: the category, a space, then the text. The words are the letters and the digits of any language (the punctuation separates the words), the same rule as in predict.

Parameters

    lang:   The language - string - required
    train_file_path:   The training file path - string - required
    iterations_param:   The iterations parameter - number (ex: 10) - required
    model_file_path_to_save:   The path of the model file to save - string - required
admin
dl n_bayesian create_model "en" "demo/train.txt" 10 "demo/model.bin";
mentdb
1
Used in use cases (3 recipes)

    Deep learning engineer › Categorise bank statement lines
    Deep learning engineer › Sentiment triage of customer reviews
    Deep learning engineer › Confusion matrix of an intent classifier
Often used with

    dl n_bayesian delete  ·  dl n_bayesian exist  ·  dl n_bayesian load  ·  dl n_bayesian predict  ·  file create  ·  file delete  ·  dl n_bayesian show  ·  json doc

dl n_bayesian load <key> <model_file_path>


Description

    To load a Naive Bayesian network model.

Parameters

    key:   The network key - string - required
    model_file_path:   The model file path - string - required
admin
dl n_bayesian load "bayesian1" "demo/model.bin";
mentdb
1
Used in use cases (3 recipes)

    Deep learning engineer › Categorise bank statement lines
    Deep learning engineer › Sentiment triage of customer reviews
    Deep learning engineer › Confusion matrix of an intent classifier
Often used with

    dl n_bayesian create_model  ·  dl n_bayesian delete  ·  dl n_bayesian exist  ·  dl n_bayesian predict  ·  file create  ·  file delete  ·  dl n_bayesian show  ·  json doc

dl n_bayesian predict <key> <sentence>


Description

    Get the category of a text with a model loaded by dl n_bayesian load: the probability of each category, the best category and its probability.

Parameters

    key:   The network key - string - required
    sentence:   The sentence - string - required
admin
dl n_bayesian predict "bayesian1" "21/06 PHIE LA BOETIE";
mentdb
{ "input": "I\u0027m happy", "prediction": "positif", "best_percent": "66,666667 %", "best_index": 0, "probabilities": [ { "prob_double": 0.6666666666666666, "index": 0, "prob_percent": "66,666667 %", "key": "positif" }, { "prob_double": 0.3333333333333333, "index": 1, "prob_percent": "33,333333 %", "key": "negatif" } ], "best_double": 0.6666666666666666 }
Used in use cases (3 recipes)

    Deep learning engineer › Categorise bank statement lines
    Deep learning engineer › Sentiment triage of customer reviews
    Deep learning engineer › Confusion matrix of an intent classifier
Often used with

    dl n_bayesian create_model  ·  dl n_bayesian delete  ·  dl n_bayesian exist  ·  dl n_bayesian load  ·  file create  ·  file delete  ·  dl n_bayesian show  ·  json doc

dl n_bayesian delete <key>


Description

    To delete a Naive Bayesian network.

Parameters

    key:   The network key - string - required
admin
dl n_bayesian delete "bayesian1";
mentdb
1
Used in use cases (3 recipes)

    Deep learning engineer › Categorise bank statement lines
    Deep learning engineer › Sentiment triage of customer reviews
    Deep learning engineer › Confusion matrix of an intent classifier
Often used with

    dl n_bayesian create_model  ·  dl n_bayesian exist  ·  dl n_bayesian load  ·  dl n_bayesian predict  ·  file create  ·  file delete  ·  dl n_bayesian show  ·  json doc