Machine Learning for Prediction

Prediction, according to Merriam Webster dictionary, is an art of declaring or indicating in advance especially foretelling on the basis of observation, experience, or scientific reason. According to Cambridge dictionary, prediction is a statement about what you think will happen in the future. Prediction is made about the outcome of future based upon a pattern of evidence. It happens based on prior knowledge or evidences. In statistics, prediction is a conclusion based on statistical inference while in science, it is a rigorous and often quantitative analysis of past and present data or occurrence to forecast what will happen under certain conditions.

Prediction has been applied in virtually every area of our life; in medical researches, engineering, geography, forecasting, finance and market, sport, games, technology, communication, construction and so on. Predictions have gone a long way into our everyday life. Amazon, Jumia, Konga predict what else you might like to buy every time you shop. Netflix and other movie sites predict the movie you might want to watch. Google is predicting how you will respond to your emails. Match.com and other dating sites are even trying to predict who you might fall in love with. We can see prediction in homes where children predict when their father will be home, wives predicting the movement of husbands. Also in the institution where a lecturer predicts the grade a student will possibly graduate with based on his current grade and his seriousness. These predictions have become part of us that we don’t always even notice them anymore.

Machine Learning has been applied to help in this prediction. Machine Learning is a current application of artificial intelligence based around the idea that we should really just be able to give machines access to data and let them learn for themselves. Machine learning can take in large dataset that human cannot comprehend and process it at a great speed. Machine Learning has been around since the 20th century but it is just finding use now because of the powerful computers we have now which are able to run it. In the 20th century, there were no powerful computer able to run it and still now, only few computers are able to run it well and efficiently. Also availability of large data enhances the use of machine learning because the algorithms used in machine need as many as possible large data to be trained with for accuracy and efficiency.

There are three methods used in machine learning: supervised, unsupervised and reinforced learning. In supervised learning, you train the algorithm with data which contains the answer. Example when you train a machine to identify your friends by name, you need to identify them for the computer. If you trained an algorithm with data where you want the machine to figure out the pattern by itself, it is called unsupervised learning. If you give a machine a goal and you expect the machine through try and error to achieve the goal, it is called reinforced learning.

Few publicized examples of machine learning applications are: the self-driving Google car, online recommendation offers such as those from Amazon and Netflix, knowing what customers are saying about you on Twitter, fraud detection, speech and image recognition.

Prediction has been so ubiquitous that we apply it to almost every area of our lives and a day cannot go without making a prediction. Prediction such as if it will rain later in the day due to the current weather condition, predicting the outcome of a football match, predicting when a person would likely get to a place based on the traffic and other logistics. Prediction makes us feel in control because when we know ahead of what will happen in future it gives us a better chance of controlling things and prepares us ahead of things to come. Prediction helps and guides our decision to achieve a goal and avoid potential discomfort. If the outcome of steps is known before taking the steps, it guides the steps to take to achieve a particular goal. Machine learning does prediction better and faster. Machine Learning can be fed with large dataset that human cannot process or data that will take human years to process which enables it to predict more accurate within a short period of time than human.