{"id":154274,"date":"2019-08-27T07:59:10","date_gmt":"2019-08-27T11:59:10","guid":{"rendered":"https:\/\/www.countingpips.com\/?p=154274"},"modified":"2019-08-27T07:46:23","modified_gmt":"2019-08-27T11:46:23","slug":"introduction-to-artificial-neural-networks-in-python","status":"publish","type":"post","link":"https:\/\/www.investmacro.com\/forex\/2019\/08\/introduction-to-artificial-neural-networks-in-python\/","title":{"rendered":"Introduction to Artificial Neural Networks in Python"},"content":{"rendered":"
August 27, 2019<\/div>
\r\n<\/div><\/div>

By Padmaja Bhagwat, Kite.com<\/a><\/strong><\/p>\n

\n
\n
\n

The Python implementation presented may be found in the Kite repository on Github<\/a>.<\/p>\n

Biology inspires the Artificial Neural Network<\/h2>\n

The Artificial Neural Network (ANN) is an attempt at modeling the information processing capabilities of the biological nervous system<\/i>. The human body is made up of trillions of cells, and the nervous system cells \u2013 called neurons \u2013 are specialized to carry \u201cmessages\u201d through an electrochemical process. The nodes in ANN are equivalent to those of our neurons, whose nodes are connected to each other by Synaptic Weights (or simply weights)<\/i> \u00a0\u2013 equivalent to the synaptic connections between axons and dendrites of the biological neuron.<\/p>\n<\/div>\n

<\/div>\n
\n

Let\u2019s think of a scenario where you\u2019re teaching a toddler how to identify different kinds of animals. You know that they can\u2019t simply identify any animal using basic characteristics like a color range and a pattern: just because an animal is within a range of colors and has black vertical stripes and a slightly elliptical shape doesn\u2019t automatically make it a tiger.<\/p>\n

Instead, you should show them many different pictures, and then teach the toddler to identify those features in the picture on their own, hopefully without much of a conscious effort. This specific ability of the human brain to identify features and memorize associations is what inspired the emergence of ANNs.<\/p>\n

What is an Artificial Neural Network?<\/h2>\n

In simple terms, an artificial neural network is a set of connected input and output units in which each connection has an associated weight. During the learning phase, the network learns by adjusting the weights in order to be able to predict the correct class label of the input tuples. Neural network learning is also referred to as connectionist learning, referencing the connections between the nodes. In order to fully understand how the artificial neural networks work, let\u2019s first look at some early design approaches.<\/p>\n

What can an Artificial Neural Network do?<\/h3>\n

Today, instead of designing a standardized solutions to general problems, we focus on providing a personalized, customized solution to specific situations. For instance, when you log in to any e-commerce website, it\u2019ll provide you with personalized product recommendations based on your previous purchase, items on your wishlist, most frequently clicked items, and so on.<\/p>\n

The platform is essentially analyzing the user\u2019s behavior pattern and then recommending the solution accordingly; solutions like these can be effectively designed using Artificial Neural Networks.<\/p>\n

ANNs have been successfully applied in wide range of domains such as:<\/p>\n