10/17/2021 0 Comments Logistic RegressionLogistic Regression, as the name of the algorithm may confuse you, is in fact a classification algorithm. Supervised machine learning is broken into two realms: regression and classification. Regression, as discussed in the previous post, involves predicting a continuous output whereas classification involves predicting a discrete output that represents a class. Many of the classification tasks that I will demonstrate in the videos deal with simply two classes. This may be distinguishing between a cancerous or benign tumor, or an iris-setosa or iris-virginica. Logistic regression models the probability that a given data point belongs to certain class; therefore, its output falls between 0 and 1. In order to do so, the logistic regression model uses the sigmoid activation function to create this binary case. Data points are projected onto the sigmoid curve using a threshold of 0.5 to decide which class a data point belongs to.
0 Comments
9/4/2021 0 Comments Multiple Linear RegressionLinear Regression is quite possibly the most widely known regression algorithm. It is the most fundamental of the other techniques (which I will cover in the future) due to its simple yet effective nature. Typically, in data science, you will face much more than just a single feature, therefore resulting in more parameters. Hence, the algorithm is commonly referred to as Multiple Linear Regression, which uses the equation below. y = ß0 + ß1x1 + ß2x2 + ß3x3 + ... + ßnxn where n is the number of features and m represents number of data values Finding the parameters requires some knowledge of Linear Algebra. The Linear Regression algorithm attempts to find what's referred to as a "Least Squares Solution", meaning that the residuals between predicted and actual points are minimized. The formula for finding a Least Squares Solution is (A^T A)x = A^T B. A is what's known as the design matrix, and is an m by n matrix that contains the parameters for each data point. x, a column vector, is what the algorithm attempts to obtain through the matrix calculations. B is the column vector of the "y-values". By left multiplying by the inverse of (A^T A) on both sides, you can determine the x column vector as (A^T A)^-1 times (A^T A) is simply the identity matrix. Linear Regression is a very robust model, but must be used appropriately in the context of a given problem. Determining which algorithm to use can be tough, but with Exploratory Data Analysis, you can generally determine which one to implement through the relation between different features. Viraj Zaveri11/30/2018 6 Comments Which AI platform do you need?AI is a vast field. We have many different platforms which have many different uses. But how many of these have a bright future? (Unfortunately, We will not be including Google, H2O.AI, and Microsoft in this article. Email us if you want an article about them.) Siri: Siri is an AI platform that is found on Apple Phones, Ipads, and Computers. Siri was acquired by Apple on October 4, 2011. Siri is very useful, as it sends emails and texts, but it is meant to remain on a smartphone, because Siri is not very advanced. Apple had released the HomePod, which was a speaker. Unfortunately, they didn't bring anything new (they were too late!) If you are wanting to make your life easier on your Iphone, Siri is perfect for that. Amazon Web Services: Now, you have probably heard of Alexa. However, Alexa is just an AI system simply for consumer products. However, AWS's capabilities are shown on the cloud platform. AWS is an affordable cloud platform, and even simple teenagers can use it for a project. It can host organizations from Khan Academy, all the way to Netflix. This is programmed in C++. If you are looking for an affordable cloud for your project, or new startup, AWS is for you. IBM Watson: IBM Watson is likely to be the most powerful AI System in the world. This is an AI system that is associated with Medicine, Business, Sports, Taxes, Oncology, etc. Watson for Health, the biggest program, focuses on assisting doctors with scanning images, handing out information needed for surgery, etc. Even though IBM Watson is very fascinating, it can only be bought by big corporations. Anay BerdeA high school student who is mesmerized by technology, especially AI. 11/24/2018 0 Comments Machine LearningMachine Learning is a very broad and significant topic that lies under the overarching branch of Artificial Intelligence. Machine Learning, in a breviloquent is a way of letting the robot determine a piece of data given past data points. The way the robot will perform this is by devising an algorithm needed to obtain that piece of information. Machine learning focuses on the development of computer programs which extract information and use that to create its own. Three of the most proverbial languages utilized for Machine Learning include JAVA, Python, and C++. A man named Alan Turning along with Marvin Minksy and Dean Edmonds first created a neural network machine which was able to learn the SNARC. The idea of ML allows you to become more efficient in your time for programming and also lets you understand the pattern behind your data. One of the major diseases that ML has helped with predictions was cancer. A group of engineers extracted information for instance gender, age group, name of hospital, and previous diseases to determine the likelihood of the next person from gaining cancer. Viraj Zaveri11/11/2018 0 Comments A World Within the Hands of Robots Oftentimes, people have conversed whether or not AI will cause a sudden increase in unemployment within the near future. Several companies have already began launching several products with the involvement of AI for instance, the Amazon Alexa and Predictive Analytics. Throughout the past few years, AI has been a major topic of competition and is rapidly increasing in its effectiveness and use. Robots have already been integrated into industrial manufacturing, however, the population is slow in considering how robots will have a major impact on the job market. Not only will robots have an effect on the job community, but they will also have a major impact on the social well-being of individuals. Robots have been placed into restaurant industries and have been utilized within self driving cars, which is deeply affecting the world in today's today and age. What's your prediction for the future of AI?
|
AuthorsAnay Berde is a high school student who is mesmerized by tech, especially AI. He is the CEO of CtrlShiftTECH and is the lead of AI. He is also the HTML Instructor. ArchivesCategories |