MultiClassClassification

Multi-Class Classification will classify the short description against labels/classes in the training data.

This engine is deployed using Docker container service. Machine learning model(Naive Bayesian Classifier) is persisted using Pickle object serialization and is deployed as a RESTful Web Service using Flask Framework, in a container.

Run this engine

Clone the git or download the content.

Create Docker image

docker build -t <image name> .
docker images

Deploy Container

docker run -name <docker name> -d -p 8500:8500 <image name>

Naive Bayesian Classifier

This is an implementation of a Naive Bayesian Classifier written in Python. The utility uses statistical methods to classify documents, based on the words that appear within them. A common application for this type of software is in email spam filters.

The utility must first be 'trained' using large numbers of pre-classified documents, during the training phase a database is populated with information about how often certain words appear in each type of document. Once training is complete, unclassified documents can be submitted to the classifier which will return label with probability.

Training

To train the utility, replace the training file and run the following the get API:

<server IP>:8500/train

For example my server IP is 10.114.220.111

0.0.0.0:8500/train
> Successfully Trained

It will return saying successfully trained.

Classification

Once training is successful, classification is performed using the following get API:

<server IP>:8500/problemclass/text="<description>"

• The text argument is the description for which you want to get the classified result.

For example:

0.0.0.0:8500/problemclass/text="Host is Unavailable"
> {"Classification":"Availability","confidence":"99.7845157"}