MindsDB and MonkeyLearn

MonkeyLearn is a No-code text analysis tool. MindsDB allows you to use pre-built & custom MonkeyLearn models to use its features like classifying text according to user needs and fields of interest like business, reviews, comments, and customer feedback.

How to bring MonkeyLearn Models to MindsDB

Before creating a model, you will need to create the ML_ENGINE for MonkeyLearn using the CREATE ML_ENGINE syntax

CREATE ML_ENGINE monkeylearn_engine
FROM monkeylearn
USING
  monkeylearn_api_key = 'monkeylearn_api_key';

Once the ML_ENGINE is created, we use the CREATE MODEL statement to bring MonkeyLearn models to MindsDB.

For this example, you will make use of MonkeyLearn’s pre-made model E-commerce Support Ticket Classifier.

CREATE MODEL mindsdb.ecommerce_ticket_classifier
PREDICT tag
USING
engine = 'monkeylearn_engine',
monkeylearn_api_key = 'api_key',
model_id = 'model_id',
input_column = 'text';

On execution, you get:

Where:

Expression	Description
ecommerce_ticket_classifier	The model name provided to the model created in MindsDB.
tag	The column that will provide the predicted result.
engine	The ML framework engine used, which is MonkeyLearn.
monkeylearn_api_key	The API Key of the model provided by MonkeyLearn.
model_id	The respective model’s ID you want to make use of.
input_column	Specifies the input column fed to the model

You can use the DESCRIBE syntax to verify the model’s status.

DESCRIBE ecommerce_ticket_classifier;

On execution, you get:

Use the SELECT statement to make a prediction on the model.

SELECT * FROM ecommerce_ticket_classifier
WHERE text = 'Where is my order? The delivery status shows shipped. When I call the delivery driver there is no response!';

On execution, you get:

Create and train a model.

You can also create a model with a dataset. For this example, we will be using a dataset consisting of messages for E-commerce support tickets. The dataset will be uploaded as a file onto the GUI.

Use the CREATE MODEL syntax:

CREATE MODEL mindsdb.ecommerce_ticket_classifier2
FROM files (select * from queries2)
PREDICT tag
USING
engine = 'monkeylearn_engine',
monkeylearn_api_key = 'api_key',
model_id = 'model_id',
input_column = 'text';

Use the SELECT statement to make a prediction

SELECT * FROM ecommerce_ticket_classifier2
WHERE text = 'I ordered 4 units but only received 3';

On execution, you get:

The MindsDB model created with the MonkeyLearn model successfully predicted the tag of an E-commerce support ticket according to the text input.