For the text classification problems the first challenge would be cleaning our data and convert that in a format which can be easily understood by the computer. Consider we have to find a genre of a book or a movie based on it’s content, the first thing we have to do is preparing the training dataset. Here we have two approaches to do that, one is a simple bag of words method and the other one is Doc2Vec. Let’s explore both the methods for predicting the movie genre based on it’s subtitle.

During automation testing or during a web scraping process, you might want to take a screenshot of the complete page. Doing so is easy. Here is the simple code snippet for the same.

For every business small or big, Human Resources plays an essential role in developing a company’s strategy and handling most employee-related activities. These are the people who work for the happiness of its employees and growth of the company. The abilities to source, engage, and retain best talents are the main catalysts in exploding the company’s success.

HTML5 contenteditable is a simple and handy option for rich text editing features on web pages. Say, for example

To increase the performance in Elasticsearch, many times we add new nodes to Elasticsearch cluster and manually change the configuration file with new IP address, etc. so that it gets added into the cluster. This takes up our time too. To make it efficient, we deceided to use AWS EC2 discovery plugin which discovers the Elasticsearch EC2 instances automatically through the host provider ‘ec2’. This plugin made our task so simple. The below changes are first tested in staging and then carefully executed in production without any issues. It takes less than an hour to install AWS EC2 plugin with the upgrade.

In normal dense network or multi layer perceptron all the neurons in one layer is connected to all neurons in the next layer, which means the weight params of the network is the multiplication product of number of neurons in the connected layers, so if we were to process the image of higher resolutions the network params will be very high and require higher computational power and won’t scale for larger images. No one wants to train millions of params for the small images.

We have already seen why convolutional neural network is suitable for image Processing. Now we can build our own image classifier using Convolutional neural network. We are implementing this using Python and Tensorflow.

When you run into an error something like this,

Kubernetes is getting popular these day because of its strong community and smoother container orchestrations.

As a forward thinking organisation, we have deployed most of our apps onto Kubernetes recently, of course with the consent and acceptance from the clients.

The simplest way to index documents in production/staging environments using Ruby on Rails framework is using the following command.

In ruby, sub and gsub are very useful and powerful string operators.