Machine Learning
The future interface of every gadget and application in the world will be machine learning enabled. The machine learning models will be responsible for deciding yes or no. Most of the time, the terms data science and machine learning are used interchangeably with each other. Data science helps evaluate the data, so it can be further used to build models in the machine learning process. The amount of data being produced every second gives us the opportunity to learn more from it. Machine learning is one of the fastest growing emerging technology from the last decade or so. The entities utilizing machine learning and making the best use out of it are on the competitive edge that saves them time and money from the use of machine learning. The vastness of machine learning can also be concluded as it is used as a research tool for finding patterns in data.
What is Machine leaning?
Machine learning is the sub domain of Data science in which we learn from data and make prediction out of it. The definition of machine learning is "The capability of Artificial Intelligence systems to learn by extracting patterns from data is known as Machine Learning." Turning daily recorded data into any given some useful piece of models with the help of computational algorithms. The real goal of machine learning is not in data itself instead to create something useful out of data using machine learning to convert data into a data product. The optimal goal for machine learning is to work fine for both test data set and training data set.
Types of Machine learning
Mainly machine learning is classified into four categories supervised, Un-supervised, Semi-supervised and reinforcement learning. The type of data and task provided determines the type of machine learning. Supervised
1) Supervised
In supervised machine learning, the target column "y" is labeled and the task is of classification, and when the target column 'y' on the data set is not labeled and is continuous the task is of regression.
Classification Algorithms: Some commonly used classification algorithm are include decision tree, xgboost, adaBoost, support vector machine, logistic regression, random forest, naive Bayes Classifier, etc. these are the most common types of machine learning algorithm used
Regression Algorithm: includes linear regression, Polynomial regression, Multiple linear, ridge and Lasso.
2)Un-Supervised
When there is no label data in the data set Unlike supervised in un-supervised machine learning the algorithm look for similarities in the data and the techniques that it uses is clustering. The clustering approach in un-supervised machine learning organizes the data with same characteristics. In unsupervised machine learning there is no labeled data therefore clustering help recognize some of the pattern.
3) Semi- Supervised
Semi-Supervised learning consist of both the types of data as input, labeled and unlabeled. Semi-Supervised machine learning is achieved by using some amount of labeled and un-labeled data during training the model.
4) Reinforcement
The reinforcement learning has no data set. In reinforcement learning the model learn from the environment using rewards and error. There is no predefined data used in machine learning. Reinforcement learning uses trail and error approach for learning, if the trying instantiate a good result than it gets reward if not than it has to try something different.
Machine learning application
There several most commonly implemented machine learning applications today. Machine learning application discuss here are those which are being used. Image processing, Robotics, Data Mining, video games, Text Analysis, Healthcare.
Image processing: The use of image recognition in terms of machine learning is different than using it for authentication purposes for still images. The image processing here shows the type of object and its movement. Image processing with the help of machine learning gives a computer human like eyes. Self-driving cars use machine learning image processing to detect objects and determine their type. The benefit of using machine learning in image processing is that it can be used for moving objects and can perform calculations.
Robotics: The most common use of robots is in the automotive industry. The implementation of machine learning makes them work more efficiently. This machine learning enables them to manage the time and speed of the production line automatically. Humanoid robots that replace housemaids are also a modern example of machine learning and robotics. The humanoid robots are built to do multiple household tasks such as vacuuming, opening the door, cleaning, and answering calls.
Video games: All modern type of vide o games make decision with the help of machine learning. On the basis of performance of the gamer machine learning determines the next stage in of the game. Today many games games uses AI agents as an opponent in the game.
Text Analysis: Machine learning in the text mining is utilized more than in any other field. Text analysis make use of the machine learning to perform sentiment analysis. Classification is done in text mining where sentiments analysis are performed. The term used for machine learning in text recognition is called NLP(Natural Language Processing).
Healthcare: In health care machine learning determines the patient is infected or not. There are many techniques of machine learning used in health care to diagnose disease. Some techniques make use of the computer vision with the help of image processing it can easily be decide whether a patient has tumor or not. Some techniques of machine learning involves patient previous history for and readings to determine whether a patient is diabetic or not.
Data analytics vs Machine learning
Analytics uses statistical programming and machine learning uses the same statistical programming as a technique to develop models and predict the results. Data analytics uses visual representation, whereas in machine learning algorithms, models and their accuracy are measured with a second set of algorithms. Data analytics is done initially as part of the pre-processing of data, and machine learning is the second stage after analytics is done. Data analytics work as evidence to determine whether the machine learning model developed is accurate or not. Machine learning is used for model building with unique characteristics; that is, the more data you feed it, the more improved it becomes.
Tags:
Data Science