Supervised vs Unsupervised vs Reinforcement

Artificial intelligence, Machine learning and Neural Network are few buzzwords in today’s world. Every body knows about it or want to know about it. This will be the trend of things going to be in next decade which will rule the technology. We can understood from this only that machine learning products capable of predicting like human. They are not absolutely correct but probabilistically best in the given condition. In other words, we can say they can imitate like an experts. You can understand where are we going to be in the upcoming years. But…

Topic modeling is a machine learning technique that automatically analyzes text data to determine cluster words for a set of documents. This is known as ‘unsupervised’ machine learning because it doesn’t require a predefined list of tags or training data that’s been previously classified by humans.

It is type of statistical model in NLP which can be used to extract abstract topic for the document. It can be used frequently and for solving different purposes. …

Natural language processing joins with neural network and allows us to read the existing document as training data. And no wonder, able to replicate the same in return once the model is trained. You can understand that we will be able to replicate style of writing.

Use Cases
Replicating the style of writing of Shakespeare or any other famous writer
Replicating the poems written in olden style. Other use cases in day-to-day work, you would have noticed while writing email / messages, getting auto suggestion based on the previous word typed by you.

Use Case : Content Generation

CG is based on content or text…

Forward and backward propagation is very important to learn machine learning. It will help you to understand inside processing of model training. Once you understood that then complexity of how machine learning works will not be riddle for me. In general, feedforward means moving forward with provided input and weights (assumed in 1st run) till the output. And, backward propagation , as a name suggests, is moving from ouput to input. In BP, we reassign weights based on the loss and then forward propagation runs. These two process are independent for the training of your model. …

Box plot is the graphical presentation of data commonly used for finding the outliers in the data. As we know, data plays very important role in machine learning end to end processing. Better the data is given to train the model, you will notice model generalizing better to unseen data. So, data is the heart to solve any problem statement. Defining a problem and collecting the data is the root to start solving the problem.

Once we are ready with the problem statement, we start collecting data from different sources. And, most of the time data is input or provided…

Machine learning is buzzword and Neural Network is the must learn thing for training the model to have better models which can generalized better with unseen data. Before NN, it is known that earlier models can overfit or underfit depends upon various conditions. Various techniques to resolve that are more data, early stopping and applying techniques to normalize the data etc. With NN, people think that they should be able to converge in each and every situations. Somehow, that is not the case and surprisingly, any NN algorithm in machine learning does not perform better due to vanishing and exploding…

Bias-Variance trade off is common term heard whenever you talk about data, machine learning and training the model using data to predict. The other big terms come out of this trade off is Underfitting and Overfitting. These are few important terms to understand and then, it can help to analyze your data and reasons behind ML model not performing upto the expectations. Let us make this terms simple and understand them. So that, in future these terms won’t trouble us much. It is one time effort and will help you to evaluate your ML models life time. …

Developing Machine Learning model, training and deploying is end to end exercise. Here, I will explain how a model can be packaged in such a way that it can be used in your python program effectively and intelligently. It is not as difficult as it should be. Once you are done with your code and your code is ready to be deployed. You will see how much flexibility it can provide you to do the things in more efficient and organized manner. Let us get our hand dirty. I am starting with the below code in setup.py …