short courses programmes

Machine Learning

Description

The today’s digital world has flooded us with huge amount of data generated from various sectors, including agriculture, telecommunication, health, and banking. In most cases, these data are used to provide inferences through which different managerial decisions can be made. For instance, we can apply traditional statistical methods on the data for the following purposes: maximize profits of our agricultural investments, provide efficient allocation of resources in telecom networks, estimate housing prices, or refine policies to enhance the socio-economic wellbeing of our communities. However, given the massive data volumes we have, statistical methods may be impractical or inefficient. In essence, some sources generate unstructured data that cannot be statistically analyzed with greater accuracy.

The advancement of technology, especially on the computational power of computers, has made it possible to address the limitations of statistical methods. Machines have been empowered to investigate and learn patterns of data, and then provide better decisions that would otherwise be practically impossible to achieve through statistical approaches. Machine learning provides us with another revealing experience of understanding structures of data and using them as the basis of making decisions. Recent developments in this field show that even complex problems can be addressed more effectively through machine learning techniques. This training gives participants tools and knowledge to address common machine learning problems.

The training gives a general understanding of machine learning concepts, including fundamental principles and tools for building real-world applications. We will design and implement algorithms that work on data encountered in our everyday lives. Furthermore, the training will cover neural networks—focusing on their practical applications in our rapidly changing technological world. You will be subjected to a hands-on practical environment that keeps learners stimulated, engaged, and challenged to come up with meaningful solutions. Therefore, this training will not just end up telling the learners what the algorithms are but also hand them tools and skills they can immediately apply after the training.

Objectives

The primary objective of this training is to equip participants with core principles and tools to address common machine learning problems. The training covers important concepts of machine learning to allow participants to build effective predictive models found in commercial and industrial products. Participants can, in addition, use the concepts to build their customized models suitable for their own applications.

Learning outcomes

Upon completion of the course, participants should have gained sufficient practical knowledge and skills to perform the following tasks:
1.	Describe opportunities and general applications of machine learning;
2.	Establish an environment to design and implement a machine learning model;
3.	Describe a machine learning workflow;
4.	Give advantages of Python over other programming languages when applied in machine learning;
5.	Illustrate how machine learning works;
6.	Design and implement machine learning models to solve real-world problems;
7.	Evaluate the performance of machine learning models;
8.	Address overfitting and underfitting problems; and
9.	Describe operations and applications of neural networks.

Topics

1.	Introduction Motivations, opportunities, and applications Machine learning tools Environment setup Machine learning workflow
2.	Python fundamentals (a) Fundamentals of Computer Programming and Introduction to Python Programming Language.	Data types and variables Program structure Program looping Introduction to numpy and pandas libraries
	(b) Hands-on practice	Practical algorithms Menu-driven Programs Mathematical problems
	(c) Solving Python programming challenges
3.	Machine learning with Scikit-Learn (a) How Machine Learning works (b) Types of learning (c) Machine Learning Vs Artificial Intelligence Vs Deep Learning (d) Specialized libraries
	(e) Machine Learning algorithms	Linear Regression Logistic Regression Multivariate classification Support Vector Machines Recommender Systems Anomaly Detection Systems
	(f) Machine Learning Techniques	Handling non-numeric data Overfitting and Underfitting Linear correlation in model development Data visualization and statistical inference techniques Model validation techniques ROC curves
	(g) Hands-on practice with a real-world dataset
4.	Introduction to Neural networks (a) Neural networks Versus human brain (b) How Artificial Neural Networks work (c) Applications of neural network (d) Basics of Deep Neural Networks

Target Group

(a) Aspiring professionals seeking their careers in machine learning,

(b) Students and analysts who want to start their Data Science career and expand their knowledge on machine learning,

(c) working professionals familiar with coding but intending to execute a machine learning role evolving around the application of datasets, and

(d) Hobbyists who wish to become Data Scientists.

Course Materials

Each registered participant will receive a copy of instructors’ slides and other materials, such as datasets and software.

Certificate of Attendance

Participants will be awarded certificates of attendance by the University of Dar es Salaam.

Fee

(a) TZS 150,000 (Tanzanians)

(b) USD 150 (Foreigners)

Payment

All payments should be made centrally through Control Numbers provided by the University of Dar es Salaam. After registration, a Control Number and invoice will be emailed to you using the information you have provided in the registration system.