Machine Learning Introduction

Machine learning (ML) is a field of computer science that studies algorithms and techniques for automating solutions to complex problems that are hard to program using conventional (Traditional) programming methods.

The traditional programming method consists of following steps:

1. Creating a detailed design for the program based on a business logic.

2. Implement the detailed design as a program in a desired computer language.

Now, this traditional software engineering approach can’t be used in a field which is so diverse and has multiple disciplines under it. Let us take some examples such as stock market prediction, image detection, automated voice recognition etc.

stock market prediction:

1.Collection of data

2.Storing it in files

3.Checking for irregularities

4. Correcting data

5.Applying statistics

6.Visualising data

7. Analysing pattern

8. Prediction using previous pattern

Image Detection:

1. Loading Image dataset

2. Reading the labelled images

3. Recognizing the image

Automated Voice Detection:

1. Loading Raw Voice Data

2. Feature extraction

3. Text output from model

4. Final transcript

Now, imagine with traditional programming skills we need to do a lot for doing all these tasks. Reading data, loading data can be done with a program. However, when it comes to slicing data, indexing, labelling, type conversion, visualising, pattern matching it needs a lot of efforts as well as workforce. Since each task is itself a big project with programming.

Moreover, a final code will be amalgamation of huge code which then slows the system or it would be impossible to do such a task as such efforts are needed for each and every problem. The best thing Software engineers has done are, they made universal programs which can be implemented on any big data problem. The term machine learning is a part of such programs. The tools which we are using it right now, such as python, R are nothing but brain child of software engineers.

Our goal is to use those tools and perform the data analysis, visualisation or to predict prices etc. without worrying what happens behind the machine learning code.

But the question is why should we use it?

Today, everyone is busy with their phone in terms of internet television, music streaming, video streaming. World is making progress in terms of internet speed, data and technology.

Information is flowing faster than light. However, have you ever imagined where it has been stored? If stored in any server or cloud then why it has been stored? Almost Every website we visit, every app we download for free collect our data. They then store it in their secure servers or clouds.

The web browser we use has cookies in it. Which keeps tracks of our browsing activities. The websites where we are visiting also posses such abilities. They take note of our behaviour of visit, buying items, ordering food and store it in their database. Such a collection of data end up in being big data.

Businesses wants to increase their profitability by selling more and more items. Machine learning helps them in achieving their goal by creating a model for them.

Recall a moment when you visited YouTube watched a video and similar video were there next time when you open YouTube.

When you buy something from amazon or flipkart , similar items or items related to such item are there in recommendation list.

When you watch a video on YouTube or on a website, you get area specific ads. How those ads appear in your browsers?

This is all due to Machine Learning. It is not just transforming business but also lives of people.

Now let us take a look of top view of Machine Learning as shown in the image given below:

1.Source and prepare your data : Data can be gathered through web scraping

2.Preparation of data : checking the correctness, removing unwanted data, outliers

3.Transform the data : Enrich data, standardize data, convert it to appropriate type

4. apply algorithm on data

5.analyze the performance and make improvements in it.

6.deploy your model : make it available in real world

7.monitor model and make changes through versions.

By going through all the above steps businesses will get the predictions and they can further decide their strategies for profit making.

Let us stop here for now, we will be studying various facets , terms and terminologies in next article in this series.