What is the First Phase of a successful ML Project? | #11 of 28 | Foundations of ML: The Big Picture скачать в хорошем качестве

What is the First Phase of a successful ML Project? | #11 of 28 | Foundations of ML: The Big Picture

Let's understand the phases of a given data sensor machine learning project starting with the Discover phase. Now, we will understand this based on the housing price prediction problem statement. Complete Machine Learning Course for FREE    • Foundation of Machine Learning (The Big Pi...   🔹 What is the First Phase of a successful Machine Learning Project 2022 To start this off, we are going to enter the discovery. So, what happens during discovery, as the name suggests, we will discover everything you need in order to solve and build the ML solution from scratch. So, in this context, what is discovering everything mean when we talk about it, we are talking about it in form of data about the algorithm that is discovered the data needed discover what sort of algorithm can that can be used and discover the various insights that is coming out of the data the information that you have collected, right. So, when it comes to data as it relates to identifying what are the different possible data sources from where you can actually collect the data, this becomes a little bit more practical now. So, you know, what sort of data is probably going to be mean that you first shortlist them and identify from what sources you will be able to collect those data, you will also want to consult with experts for what data is going to be useful. You will also want to identify what is the critical data in order to solve your problem. Now, if you want to collect data, you can keep on collecting a lot of different information about the houses, but there is certain base core information without which predicting the price of the house might not be realistic, right. For example, the area of the house is a very basic information. So that will definitely be a critical data location of the house. So such data you might want to classify it as definitely I will need this data in order to make this prediction. So you call it as critical data, there are certain other data that can add more nuances to your prediction. That is it might be able to increase the accuracy of your prediction for certain types of house units. Those can be additional augmented data. Right. So that is the information about data that you want to discover. Now coming to algorithms, before you start building your machine learning model, you can of course, you can just randomly pick an algorithm and start building it. But before you go into where you can go a data organized or structured fashion. To do that, what are steps you can undertake is read the literature on existing solutions that have already been implemented or a similar problem. Similar or exactly the same problem. If it is available, you don't have to go with the exact same approach proposed in the literature. But it is a good idea to understand what our work has been done previously on such a situation. And as you go about learning the literature and finding out more about things, what you can do also understand what are different advantages and drawbacks of the various techniques that you are analyzing. You can always look this up from technical articles, research work, GitHub repositories, blog posts from various different blogs, you can even ask people around in your organization, or simply consult more senior folks in your team who has been around doing similar work for a longer period of time or similar work in a different project or vertical assets. So now you have collected the data you have you have analyzed or formed the different data sources and possibly collected the data you also have the algorithms, and then your background work on the literature. Let me know in the comments section if you have any questions! If you enjoyed this video, be sure to throw it a like and make sure to subscribe to not miss any future videos! Thanks for watching! #machinelearningplus #python #machinelearning #datascience