# Machine Learning [[Book - Machine Learning for Dummies]] [[Data Analytics Begins with Me - NUS Course.pdf]] #machine_learning ## **What is Machine Learning?** Machine learning is an application of artificial intelligence that helps AI systems learn and improve from experience. Successful machine learning training makes programs or AI solutions more useful by allowing them to complete their work faster and generate more accurate results. The process of machine learning works by forcing the system to run through its task over and over again, giving it access to larger data sets and allowing it to identify patterns in that data, all without being explicitly programmed to become “smarter.” As the algorithm gains access to larger and more complex sets of data, the number of samples for learning increases, and the system is able to discover new patterns that help it become more efficient and more effective. ## **How Does Machine Learning Really Work?** The process of machine learning relies on two different types of learning, called Supervised Learning and Unsupervised Learning. ## **Supervised Learning** Supervised learning is a process that trains the system on known input and output data so that the system can do a better job of predicting future outputs. To put it a little more simply, supervised learning requires that someone is in charge of providing feedback to the AI system, training the system to make the right decisions by labeling the data. Basically, supervised learning shows the system what conclusions it should arrive at by showing it previous sets of data, and the conclusions it should have arrived at based on that data. This helps train the system to look for data patterns, interpret those patterns, and calculate the correct answer, based on what’s worked previously. Supervised learning is typically used when the system needs to make a prediction, like when the system is tasked with estimating house prices, or determining if a picture is a cat or a dog. **Examples of supervised learning include:** - Predicting House Pricing - The system is given a whole series of input data points, like square footage, the number of bedrooms and bathrooms, features of the house, along with an output data point, the value of the house, and the system learns to predict any new house’s price based on patterns in the previous data sets (i.e. more bedrooms means a higher price). - Image Recognition - The system is shown pictures of cats and dogs, with labels assigned to each image so that it can learn which types of pictures and patterns in pictures represent a cat, and which represent a dog. The system can then be shown new images of cats or dogs, and use its pattern recognition “experience” whether the new image shows a picture of a cat or a dog. ## **Unsupervised Learning** Unsupervised learning is a process that trains an AI system to find hidden patterns or intrinsic structures in input data, without regard to outputs. In this way, unsupervised learning lets the AI system draw inferences directly from data fed into the system. Unsupervised learning is typically used for problems that require exploring data and looking for internal representations within the data, or what machine learning specialists call “clustering”. Clustering is the process of automatically grouping together different points of data that feature similar characteristics, and assigning them to “clusters.” **Examples of unsupervised learning (really “clustering”) include:** - Customer segmentation - Identifying particular customer groups that should be targeted via different marketing strategies. - Recommendation systems - Netflix’s suggestions for what to watch next, or Amazon’s suggestions for what to buy next, based on grouping together users that had similar viewing or purchasing patterns. - [[Anomaly detection]] - Banks attempting to detect fraudulent financial transactions or Airlines trying to detect defects in mechanical parts.