Hello and welcome to the second course in accounting 570. Today, I'm standing in the atrium at the Beckman Institute, which was opened in 1989. The Beckman Institute for Advanced Science and Technology was funded by a generous gift from Arnold and Mabel Beckman. Arnold was a 1923 graduate from the University of Illinois, and later founded Beckman institutes, a company that made scientific equipment. The institute forms part of the northern boundary of the University of Illinois campus, has six floors and over 300,000 square feet of office and laboratory space. This institute focuses on interdisciplinary research, which currently focuses with around three themes: integrative imaging, intelligence systems, a molecular and electronic nanostructures. Researchers from around campus worked together at the institute, and many of their efforts involve research into the development or application of machine learning. For example, work is done to aid in language translation or understanding from audio data to enable automated identification of objects in imaging data, such as animals seen by sensors in the wild, to areas of concern and medical images and to model physical systems at the nanoscale. In this course, we want to learn the basic concepts used by these groups who apply machine learning to better understand the physical world. In our case however, the goal is to understand how to make sense of the wealth of data generated by companies. By doing this, we will be able to use these data to understand the health of a company and to make future predictions on the performance of a company. By completing the data analytics track, you will understand the power of machine learning and therefore, you will be able to leverage this new technology to position yourself ahead of the competition. This course is the second in a four course sequence. Two courses comprise accounting 570, and two comprise accounting 571, which will cover data and statistical analysis for accountancy. This course will introduce machine learning, where data are used to generate a model that can be used to either make predictions or to better understand the processes that generated the data itself. This course will cover different types of machine learning such as supervise and an unsupervised learning, specific machine learning algorithms such as linear and logistic regression, nearest neighbor algorithms, naive Bayes, decision trees, and support vector machines. The course will also cover the process of moving machine learning from an analyst to a production environment, which includes the development and application of machine learning pipelines. We'll introduce ensemble learning, where the predictions of many weak learners are combined to generate more accurate and robust predictions. We'll introduce the perils of overfitting and techniques such as cross validation and regularization that reduce the likelihood of this occurring. We'll also look at the careful selection of the most important existing features or the creation of new features to improve the results of machine learning, a process known as feature engineering. We'll look at the identification of clusters of data points by using algorithms, such as K-means or DBSCAN, and the identification of anomalies either by using statistical or machine learning techniques. This can lead to the identification of novel events or potential fraud. This particular course consists of eight modules. Each module contains videos, readings, and interactive notebooks. My advice for this second course remains the same as for the first. To make the most of this course, complete the readings, run the interactive notebooks, and try the student exercises contained therein. And finally, be sure to ask questions in the class forums. Before concluding, I would like to remind you of the quote from Rhett Allain I shared at the start of the first course, "Confusion is the sweat of learning." I expect you will learn a lot in this course, but it will be challenging. I hope that you find this course to be a rewarding and enjoyable experience. Good luck.
