Machine learning algorithms play a pivotal role in today's data-driven world. They enable us to extract meaningful insights, make accurate predictions, and automate complex tasks. While there are numerous libraries and pre-built models available, it is important to understand and implement machine learning algorithms from scratch. Doing so allows us to delve deep into the underlying principles and mechanisms, fostering a comprehensive understanding of how these algorithms work. By implementing them from scratch, we gain insights into the intricate details of data preprocessing, feature engineering, model training, and evaluation. This hands-on experience helps us grasp the nuances and challenges associated with different algorithms, enabling us to tailor them to specific problem domains and make informed decisions. Moreover, implementing algorithms from scratch promotes creativity, as it allows for experimentation and customization, opening doors to novel techniques and improvements.

Implementing machine learning algorithms from scratch provides a solid foundation for aspiring data scientists and machine learning practitioners. It enhances our problem-solving skills, as we gain a deeper understanding of the underlying mathematics and concepts behind these algorithms. By writing code for each step of the algorithm, from data preprocessing to model evaluation, we become proficient in manipulating and transforming data, developing a strong intuition for handling real-world datasets. Moreover, implementing algorithms from scratch builds resilience and adaptability. It empowers us to overcome challenges and troubleshoot issues that arise during the implementation process. Understanding the inner workings of these algorithms also enables us to better interpret their results and make informed decisions based on the model's predictions. Ultimately, by implementing machine learning algorithms from scratch, we not only develop technical expertise but also foster a deep appreciation for the intricacies and power of these algorithms, allowing us to become more effective and innovative practitioners in the field of machine learning.

Algorithms:

• Linear Discrimanant Analysis link
• Logistic Regression link
• K-Means Clustering link
• Artifical Neural Network link

There is a detailed readme file for each implementation. The code contains comments too to improve readability.

Machine Learning: A Probabilistic Perspective by Kevin P. Murphy