Introduction to Predictive Analytics Models

A free video tutorial from Packt Publishing
Tech Knowledge in Motion
Rating: 3.9 out of 5Instructor rating
710 courses
420,908 students
Introduction to Predictive Analytics Models

Lecture description

Present an overview of the section. Discuss the concepts of Predictive Analytics and its relationship with Machine Learning and give some characteristics of ML models.

Learn more from the full course

Become a Python Data Analyst

Take your data analytics and predictive modeling skills to the next level using the popular tools and libraries in Pytho

04:29:56 of on-demand video • Updated June 2017

Learn about the most important libraries for doing Data Science with Python and how they can be easily installed with the Anaconda distribution.
Understand the basics of Numpy which is the foundation of all the other analytical tools in Python.
Produce informative, useful and beautiful visualizations for analyzing data.
Analyze, answer questions and derive conclusions from real world data sets using the Pandas library.
Perform common statistical calculations and use the results to reach conclusions about the data.
Learn how to build predictive models and understand the principles of Predictive Analytics
English [Auto]
Welcome to the last section of this course introduction to predictive analytics models let's see what we are going to learn in this section. First we are going to talk about what is predictive analytics then we will talk about machine learning which is one of the most successful tools for doing predictive analytics. After that we will introduce the psychic learn library which is the main library in Python for doing machine learning. And finally we will use the data sets from previous sections to build a classification in a regression model and we will use these models to make predictions. This is the first video in D.S. introduction to predictive analytics in this video. Well we'll answer the question what is predictive analytics. We will talk about the machine learning approach to predictive analytics. We will briefly talk about the types of machine learning models. And finally we will see what are the components of supervised learning model. OK so let's talk about predictions and the first thing I want to clarify is that in this context prediction does not necessarily refers to the future. When I say prediction we mean it's a guess for something that is not known something that we have not observed yet. OK so how do we make predictions. Well there are many ways to the predictions you can for instance ask a high priest or a psychic and this is what we humans kept on for millenia or their approach could be two years. Your intuition or it was unexpected. And these are more traditional business practices any more situations. This is the only possible way to do it. And of course are there ways and there is relatively new and very successful approach called predictive analytics in its use has increased dramatically in recent years because of two things. First we have now the means to do it which our technology and data. And second it works really really well. Predictive analytics is the use of data combined with techniques from mathematics is that they extend computer science to make predictions about long known events. The goal of predictive analytics is to produce a good assessment of what may happen with an unknown event. So how to do predictive analytics Well there are many ways to do predictive analytics. But one of the most successful tools for doing predictive analytics is machine learning. So now let's talk about machine learning machine learning is a soft field of computer science and in general can be simply described as giving the computer the ability to learn without being explicitly programmed. The field of machine learning has the butt of many methods to teach computers to perform certain tasks using data. And this approach has been very successful for doing predictive analytics. Now let's talk about the types of machine learning problems that we can have. We can separate learning problems in a few large categories and for our purposes we will consider just two groups. In fact in this section we only talk about supervised learning which has two main subtypes of problems classification integration problems. I just wanted to let you know that there is another big branch of pushing you are being called on supervised learning. But but we won't be talking about that branch in this section. Now let's see what are the elements that we need in order to do a supervisor. How do we know that we shouldn't use supervised learning. Well the key is this is supervised learning. We have samples or observations about something and for each of cerebration we have different features also called attributes or variables in a Target variable that we will like to predict. So the target variable is the one that we would like to predict. For instance here I am showing a few rows of the students data set that we used in previous sections. And as you can see for every of salvation every observation represents a student and for every student we have some features like gender age family size and so on. And if we define one of these features as the target the one that we would like to predict then we are talking about supervisory OK. As I said before we have two main types of supervised learning problems and the difference between them is very simple. You know that you have a classification problem if the target variable is a categorical variable. On the other hand if the target variable is a numerical variable then we are talking about a regression problem for each of these class of problems. There are many models that we can use. Finally I would like to make two important points about this section. The first one is that we will treat the models in this section as a black box. So I'm not going to explain any details about how they work internally. This is a high level introduction and I will just show you the big picture on how to build predictive models. But of course machine learning is a huge topic and we don't have time to this cause. The most basic concept let alone deeper topics. So just the big picture here OK in this video we learned about analytics and we learned some concept about supervised the machine learning in the next video. We will talk about the Sikander in library and we will build a simple predictive model.