Machine Learning for Beginners

A Plain English Introduction to Artificial Intelligence and Machine Learning

John Slavio

• Publisher: Abhishek Kumar
• eBook ISBN: 9781975679941
• Topic: Computer Science
• Subtopic: Artificial Intelligence (AI) & Semantics
• Index: Computer Science

So you've heard a lot about AI and Machine learning. But what does it actually mean?
This book could be the answer you're looking for... This book is an introduction to basic machine learning and artificial intelligence. It gives you a list of applications, and also a few examples of the different types of machine learning. Here's What You'll Learn in this Book:

Introduction to Machine LearningDifferent Applications of Machine LearningIntroduction to Statistics for Machine LearningSupervised LearningUnsupervised LearningReinforced LearningConclusion

SUPERVISED MACHINE LEARNING ALGORITHMS

Supervised machine learning algorithms are basically machine learning algorithms where you give the machine a helping hand in the beginning. These usually lead to unsupervised machine learning or even reinforced machine learning. Whenever you perform supervised machine learning, you are handing the machine the data sets and setting it up so that they get the right answer almost all the time. Once they get an incorrect answer, you then have to go back and see why it got the incorrect answer. Supervised machine learning is the first stage to almost all forms of machine learning because you have to watch and ensure that what is coming out will be correct or mostly accurate when you don't watch it or help it. This is why we need to change our algorithm to handle large sets of data that will have a helping hand in creating bias. Essentially, what we are going to do here is we are going to hand the machine learning algorithm (that we've partially created) the most biased answer we can possibly give it. However, we first have to rewrite our algorithm to handle years as one chunk of data. It is unlikely that students would have their ethnicity held in one database and their field of study in another database, so we need to change our algorithm to handle a multidimensional array.

var yearOne =[[][][]] var yearTwo =[[][][]] var yearThree =[[][][]] // 1,2,3,4,5,6,7,8,9,10 var commonEthnicity = [0,0,0,0,0,0,0,0,0,0] var commonOrigins = [0,0] var commonStudy = [0,0,0,0,0,0] function machineLearning(year){ var ethnicity = year[0] var origin = year[1] var study = year[2]

Now that we have set it up to handle multidimensional arrays, we need to create a bias in the data. The usual method for doing this is to make the first data set completely biased, the second data set partially biased, and the last data set with only a slight bias. This gives the machine a starting point and then slowly balances the bias out so that when it hits random numbers, it’s able to possibly make some more accurate guesses. In reality, you would have a large database where you would use testing sets to test your algorithm and once your algorithm showed promise, you would use the testing data instead of the multidimensional arrays we’ll be using here. Then you would use your algorithm on the remaining amount of data to reap the benefits of non-simulated machine learning. Let’s lay some ground rules for our experiment:

1. The end of the third testing stage will be a 2, 1, and 4
2. Each year will hold 50 students
3. The last year will be randomized but also doctored to provide bias.

Now, here is what the first year looks like:

var yearOne =[[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2],[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1],[4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4]] var yearTwo =[[],[],[]] var yearThree =[[],[],[]]

And this is our result:

It is likely that the next student will be a 2 with a 1 and they will study the field of 4.

As expected, we have a completely biased answer. Now we need a multidimensional array that’s only partially biased. Let’s say our partial will be 50%. This is what the addition of the second year looks like:

var yearOne =[[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2],[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1],[4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4]] var yearTwo =[[2,2,2,10,2,7,2,5,2,3,2,9,2,10,2,1,2,5,2,3,2,10,2,1,2,10,2,7,2,5,2,8,2,4,2,8,2,4,2,2,2,6,2,6,2,4,2,2,2,4],[2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1],[4,2,4,6,4,5,4,2,4,4,4,6,4,4,4,4,4,2,4,6,4,6,4,6,4,5,4,6,4,4,4,5,4,4,4,2,4,4,4,2,4,4,4,6,4,4,4,5,4,6]] var yearThree =[[],[],[]]

And this is our result:

It is likely that the next student will be a 2 with a 1 and they will study the field of 4.

As you can see, our numbers have changed very little in between the years. Now it is time for the last year and only 10% of it will be biased.

var yearOne =[[2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2,2],[1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1],[4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4,4]] var yearTwo =[[2,2,2,10,2,7,2,5,2,3,2,9,2,10,2,1,2,5,2,3,2,10,2,1,2,10,2,7,2,5,2,8,2,4,2,8,2,4,2,2,2,6,2,6,2,4,2,2,2,4],[2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1,2,1],[4,2,4,6,4,5,4,2,4,4,4,6,4,4,4,4,4,2,4,6,4,6,4,6,4,5,4,6,4,4,4,5,4,4,4,2,4,4,4,2,4,4,4,6,4,4,4,5,4,6]] var yearThree =[[2,3,6,1,3,2,3,10,7,10,2,8,6,7,8,2,9,3,3,7,2,3,9,6,4,2,3,8,6,8,2,4,7,2,9,2,9,1,9,6,2,4,6,10,4,2,7,10,1,1],[1,2,2,2,1,1,1,2,2,1,1,2,1,1,1,1,1,1,2,2,1,2,1,1,2,1,1,2,2,1,1,1,1,1,1,1,1,2,1,1,1,1,1,2,1,1,2,2,2,1],[4,6,4,2,5,4,1,3,6,4,4,4,6,4,2,4,2,6,2,1,4,1,6,3,5,4,4,1,2,2,4,4,6,3,3,4,2,5,6,3,4,2,6,5,6,4,2,6,6,6]]

And this is our result:

It is likely that the next student will be a 2 with a 1 and they will study the field of 4.

Alright, so even with a bias of 10%, we can still clearly see that we are getting the results that we want. However, we’re still missing two very important components: Confidence level and Final Test. In order to figure out the confidence level, we need to be able to calculate the percentage of each category and then combine those into a single percentage. Since we chose an easy sample party of 50, this means we just double the number to 100 and the number itself. For instance, if there were 14 1’s out of 50, we would say that there is a 28% confidence there is going to be a 1. Therefore, after some rearranging and rewriting, our code looks like this.

function machineLearning(year){ var ethnicity = year[0] var origin = year[1] var study = year[2] var commonEthReal, commonOrReal, commonStReal; var commonEthPt, commonOrPt, commonStPt; var confidence; var placeholder; var i; for(i = 0; i < ethnicity.length; i++){ commonEthnicity[ethnicity[i] - 1]++ } commonEthPt = Math.max(...commonEthnicity) commonEthReal = commonEthnicity.indexOf(commonEthPt) + 1 for(i = 0; i < origin.length; i++){ commonOrigins[(origin[i] < 2 ? 0 : 1)]++ } commonOrPt = Math.max(...commonOrigins) commonOrReal = commonOrigins.indexOf(commonOrPt) + 1 for(i = 0; i < study.length; i++){ commonStudy[study[i] - 1]++ } commonStPt = Math.max(...commonStudy) commonStReal = commonStudy.indexOf(commonStPt) + 1 confidence = ((commonEthPt + commonOrPt + commonStPt) / 3) placeholder = "It is likely that the next student will be a " + (commonEthReal) + " with a " + (commonOrReal) + " and they will study the field of " + (commonStReal) + ". I am " + confidence + "% sure that this will happen." console.log(placeholder) }

This looks very nice and now when we run it on the 50% biased year, we receive this:

It is likely that the next student will be a 2 with a 1 and they will study the field of 4. I am 50% sure that this will happen.

As you can see, the confidence level is exactly 50% sure of this. Let’s see what happens with year 2 and 3.

It is likely that the next student will be a 2 with a 1 and they will study the field of 4. I am 28.666666666666668% sure that this will happen. It is likely that the next student will be a 2 with a 1 and they will study the field of 4. I am 20% sure that this will happen.

Those are some extremely intriguing results, but we’re still not quite done bec...

Table of contents

• DISCLAIMER
• ABOUT THE AUTHOR
• WHAT IS MACHINE LEARNING?
• IMPORTANCE OF MACHINE LEARNING / APPLICATIONS
• INTRODUCTION TO STATISTICS FOR MACHINE LEARNING
• SUPERVISED MACHINE LEARNING ALGORITHMS
• UNSUPERVISED MACHINE LEARNING ALGORITHMS
• REINFORCED MACHINE LEARNING ALGORITHMS
• CONCLUSION