Chapter 4. Basics of Machine Learning
Any form of any analytical activity depends heavily on the presence of some clues or data. In today's world data is bountifully available. Due to the broad availability of various devices (such as mobile devices), IoT devices, or social network, the amount of data generated day by day is exploding. This data is not all waste; it can be used to make lots of deductions. For example, we can use this data to figure out what particular ad the user might click on next or what item the user might like to purchase along with the item they are already purchasing currently. This data can help us figure out a knowledge base that can directly impact the core business in many useful ways, hence it is very important.
This chapter is action-packed and we will try to cover a lot of ground while learning the basics. In this chapter, we will cover:
What is machine learning?
Machine learning is a form of artificial intelligence where a computer program learns from the data it is fed or trained with. After learning from this data it internally builds a knowledge base of rules and based on this knowledge base it can later make predictions when it is fed new data. Machine learning is part AI, part data mining, and part statistics, but overall the criterion is to teach a machine to make new decisions based on past data it is trained with. So, for example, if we teach a machine some data regarding the inventory statistics of a store throughout the year then you might be able to tell things such as in which months the items sell more or which items sell more often. Also, it can tell the shop owner if they are selling one particular item more than other items; it can also show this to the customer so as to increase sales.
The concept of making new predictions is very important as we can now make predictions such as in which zone or area a marketing...
This chapter was action packed on machine learning and its various concepts. We covered a lot of theoretical ground in this chapter by learning what machine learning is, some important real-life use cases, types of machine learning, and the important concepts of machine learning such as how we extract and select features, training our models, selecting our models, and tuning them for performance by using techniques such as training/test set and cross validation. We also learnt how we can run our machine learning models specifically on big data and what Spark has to offer on the machine learning side in terms of an API.
In the next chapter, we will dive into actual machine learning algorithms and we will learn a simple yet powerful and popular linear regression algorithm. We will understand it by using an example case study. After studying linear regression we will study another algorithm logistic regression and we will also try to learn it by using a sample case study.
