The Anatomy Of Data?Science
One of the main goals of data science is to predict outcomes based on incoming data. The greater variety in the samples you have, the easier it is to find relevant patterns and have a more accurate prediction of the result. This is the reason why several businesses like Google create systems like ReCaptcha in order to gain a large variety of data.
Components of?ML
In order for us to do all of that we use Machine Learning which is using computer systems that are able to learn and can adapt without needing explicit instructions. Within Machine Learning there are 3 components needed for it to be done successfully: data, feature and algorithm.
To begin with, Data is a unit of information, the more relevant data you have the better the result. This can be anything from how many customers you get to what time your customers show up and which employee has the least customer satisfaction score. There are two main ways to get the data: manual and automatic. Manually collected data contains less mistakes but takes longer to collect making it more expensive. On the other hand the automated data collection approach is cheaper but in return you will be gathering everything you can find and hoping for the best. It is extremely tough to get a good collection of data mainly because many businesses refuse to reveal their dataset and everyone else has to start collecting from scratch.
Features (also known as parameters or variables) are the next thing we’ll talk about. These could be anything from your customers age, their gender, how often they come to make a purchase, which item is used the most etc. These are what differentiate the different data and what a model looks at when making a prediction. If the data is in tables these would be the column names. With there being several types of data there are several features that are hard to categorise. This is why selecting the right features is usually the longest part of the process. Unfortunately this is also the main source of mistakes since humans are subjective and an objective eye is really needed.
领英推荐
Moving on to Algorithms which, on the other hand, is the method you choose to analyse and solve the data to make the prediction. The algorithm used can affect the precision, performance, and size of the final model. While the algorithm of a model is important, what is important is data: if the data that is collected is not good, even the best algorithm won’t help. This means that the percentage of accuracy of the algorithm isn’t important at first so trying to acquire the data is the most important step.
All of this helps in this are the building blocks for Data science and Artificial intelligence.
To prevent any faulty linear regression let’s look at what exactly Artificial intelligence is and the building blocks that make it up.
What is Artificial Intelligence?
Artificial intelligence is the umbrella name that covers the whole knowledge field and is the development of computer systems able to do tasks that require human systems. Going down we have Machine Learning which is a part of artificial intelligence, an important part, but not the only one. Machine learning is the development of computer systems that are able to learn and can adapt without needing explicit instructions.
Scrolling a little lower we have Neural Networks. Neural Networks are computer systems that are modelled on the human brain and nervous system. These are one of the many machine learning types, there are the more popular aspects in the data science and AI field but there are other aspects as well.
Then you have Deep Learning which is a modern method of building, training, and using neural networks. It is basically a new architecture. Nowadays in practice, no one separates deep learning from neural networks since most even use the same libraries for them.
With all of this information I hope that this helps you understand a little more. For those of you who are worried about us accidentally creating Skynet, we’re nowhere near that… yet.