What you’ll do
This diploma course in machine learning with Python highlights the many benefits that Python has within the frameworks of artificial intelligence (AI). Machine learning is a specific branch of artificial learning that is associated with computer science. It is about making software applications predict outcomes more accurately without any prior explicit programming. It provides computers with the capacity to learn without being obviously programmed. This course will cover two types of machine learning: supervised and unsupervised. Linear regression used to solve a regression problem is an excellent example of supervised machine learning. On the other hand, highlighting various customer groups to strategise marketing is a practical example of unsupervised learning. Within these two types of machine learning, various models depend on the purpose you would like to achieve. For instance, there are parametric and non-parametric models, as covered in this course.
As you work through the content, you will see that you resort to a parametric model when you precisely know which model you will fit to the data. Conversely, in a non-parametric model, the data directs you to what the regression should look like. For instance, the key nearest neighbours (KNN) is a non-parametric method that you use for classification and regression. The KNN has led to a wider variety of applications, particularly in text mining, agriculture and finance. As with any modelling system, the KNN has its pros and cons. You will notice that the KNN is very easy to explain, simple to understand, and extremely powerful. Furthermore, KNN does not require any assumptions on the data distribution nor request any prior knowledge. However, the efficiency of the algorithm declines very quickly as the dataset grows. Moreover, KNN cannot work if there are any missing values.
Another crucial theme covered is the use of algorithms. Many algorithms will be examined, each of which, as models, has its advantages and disadvantages. In this course, you are taught about the random forest. It is a slightly modified bagging algorithm, a collection of a decorrelated decision tree. The random forest is used because it works well with non-linear data, lowers the risk of overfitting and runs efficiently on large datasets. Conversely, it is biased while dealing with categorical variables. Thus, it is far from being suitable for linear methods with a lot of sparse features. Other methods, such as models and related algorithms, using Python as the backdrop programming language, are also delved into. This course is a windfall to computer science students and professionals who are already familiar with Python and would like to become an expert in the field of machine learning. This course has not been updated with the use of Generative AI models, like ChatGPT.