Logistic Regression

Machine learning is a science that allows a computer to take action without explicit programming. Over the past decade, machine learning has brought us an understanding of automatic driving, practical speech recognition, effective web search, and significant improvement of the human genome. Machine learning is currently very common and you can use it dozens of times a day without knowing it. Many researchers also believe this is the best way to advance human AI. In this course you will learn the most effective machine learning skills, practice practicing them, and make them work for themselves. More importantly, you learn not only the foundation of theoretical learning, but also the practical skills necessary to quickly and powerfully apply these technologies to new problems. Finally, as machine learning and artificial intelligence are related, we will learn about some of the best practices in Silicon Valley innovation. In this course, we will introduce a wide range of machine learning, data mining, and statistical pattern recognition. (I) supervised learning (parametric / nonparametric algorithm, support vector machine, kernel, neural network). (Ii) unsupervised learning (clustering, dimension reduction, recommendation system, deep learning). (Iii) Best practice in machine learning (bias / dispersion theory; machine learning and innovative process in AI). This course also covers a number of case studies to learn how to apply learning algorithms to construct intelligent robots (perception, control), text understanding (web search, anti-spam), computer vision, medical informatics, speech And draw lessons from applications. , Database mining, other fields

Logistic regression is a way of classifying data into individual results. For example, you may use logical regression to classify email as spam or non-spam. This module introduces the concept of classification, the cost function of logistic regression, and the application of logistic regression in multiclass classification.

If the dependent variable is binary (binary), logistic regression is performed using appropriate regression analysis. As with all regression analysis, logistic regression is predictive analysis. Logistic regression is used to describe data and to describe the relationship between dependent binary variables and one or more independent, sequence, interval, or proportional level independent variables. Logistic regression can test the type of problem. In discriminant analysis, two or more groups or clusters or groups are known in advance and one or more new observations are classified as one known based on the measured characteristics. Group discriminant analysis models the distribution of predictor X for each response category and then uses Bayes' theorem to convert these theorems into estimates of the response class probability for a particular X value. These models can be linear or quadratic

What is the difference between logistic regression and linear regression? Logistic regression gives discrete results, but linear regression gives continuous results. A good example of continuous results is a model to predict the value of a house. This value is always different depending on parameters such as size and position. A discrete result is always one thing (you have cancer) or another thing (you do not have cancer). As with linear regression, logistic regression is more efficient by deleting attributes that are very similar (related) to attributes that are not related to output variables. Therefore, feature engineering plays an important role in the execution of logistic regression and linear regression. Another advantage of logistic regression is that implementation is very easy and can be trained very effectively.