Bayes Theorem in the World of Data
This article discusses how Bayes Theorem is being used in the world today and how big data is a major contributing factor in its increased use. Bayes Theorem employs a method of both predicting the future and classifying large amounts of data. According to the article, Bayes Theorem is important because, “You can classify the most likely outcome of a future event based on historical data from the past.” The article continues by describing how a Bayes Theorem based program or “classifier” is created and how it must first be trained to accomplish its goals. This can be done by using previous data, of which there truly is an abundance according to the article. One such example of this type of data is historical data on weather and weather patterns. Some of this data can be fed to a robot, and then its predictions can be compared with the actual outcomes of the following year. For example, the program can be provided with data from 2005 to predict the weather in 2006. Then, the results of what actually happened in 2006 can be compared with the program’s results. The machine then learns and is trained to make more accurate predictions. According to the article, “The more data that is used to train the Bayes Classifier, the more accurate it will become over time.” This is a pure example of a machine learning and then adapting to previous results. Overall, it is clear that Bayes Theorem when applied to the real world using big data can be used to understand “current customer sentiment, potential customer actions, and many other types of observational data either at rest or in motion.”
In class this week we learned about Bayes Theorem and discussed many of its practical uses. We were introduced to the equation for Bayes Theorem which is P(A|B) = (P(A)*P(B|A)) / P(B). This formula means calculated the probability of event A occurring given that event B has already occurred. As an example, in class we reviewed the probability of a restaurant being good in reality, after a good review is read by a potential customer. Another interesting example covered in the article is the probability that someone is going to stop using an app depending what IOS system they are using. This is important because it may show that there be a bug in the app when it is being used by a certain operating system. It is quite clear that Bayes Theorem will continue to have a major impact on a world, like ours, where the amount of available data is extremely large. It will be quite fascinating to see the new applications developed that take advantage of Bayes Theorem.