Aman_ML_Machine Learning Introduction.md

What is Machine Learning?

Machine Learning has become a buzzword in the field of technology nowadays. There is race among enterprises to add functionalities in their operations by using machine learning. This has also increased the opportunities for aspiring data scientists. So having knowledge of machine learning is crucial at this moment for every tech enthusiast.

Machine Learning in simple words is teaching computers to learn by themselves without explicitly programming them so that they can perform various repetitive tasks by themselves. Since machines never get tired like humans they can do tasks for longer time. Machines learn when we teach them the pattern of doing any work. They start learning by themselves when they start recoignizing the pattern.

For example, the task of filling water in the tank whenever the tank gets empty. If we teach the machine to detect whether the tank is empty or full and start the motor whenever the tank is empty, it would do this repetitive task on its own. Similarly we can also teach it to stop the motor whenever the tank is full by detecting the water level in the tank.

So the aim of Machine Learning is enabling machines to learn by themselves. Machines need a machine learning model to learn. Data is collected and fed into the machine learning model so that machines can recognize the pattern and do the work on their own.

Steps for preparing a machine learning model are

1. Collection of Data

   We need to collect the right data according to the problem we want to solve. Different problems have different types of data. We can search for the relevant datasets from sources like Google Datasets Search, UCI Machine Learning Repository and other external repositories. `

2. Handling the missing data

   After the collection of data, it is necessary to assess the condition of the data. Since the collected data is rarely perfect dataset, we need to look for incorrect, inconsistent or missing data.

3. Making the data consistent

   The data should be formatted so that it fits our machine learning model. There is irregualrity in the data when it is collected from different sources. There can also be aberration in the datasets if it has been collected by more than one person. There is decrease in errors when the data is formatted consistently.

4. Deciding important key factors

   It is the job of our model to decide which feature is important and which is not on the basis of its affect on the output. We can simply provide the data sets and let the machine do it's job.But doing this consumes more power and time. So we can help our model by not feeding the features that are unrelated to the output.

5. Splitting data into training and testing sets

   This is the final step in preparing our model that involves splitting our data into two sets. The first set of data to train our algorithm and another set for testing. The selected data for training and testing should not be overlapping subsets of our dataset.

In this way we build our machine learning model by following the above mentioned steps. Once the model is built, we deploy the models.

Why Machine Learning?

Applications of Machine Learning

Machine Learning has numerous applications in our daily life. We see the use of machine learning everyday but are not aware of it. Some examples where machine learning is applied are:

1. Face detection and finger print detection in our smartphones for unlocking our devices
2. Recommendations in social media sites on the basis of our browsing history
3. Fraud transactions detection system used by banks
4. Netflix movies recommendation on the basis of our watch history

These are only some few examples of use of machine learning models that we encounter daily. There are many such examples where machine learning is applied.

🐦 twitter | 📷 instagram | 👔 linkedin