Machine learning --Introduction.pptx
- 2. Learning objectives for this section Having completed this section successfully, you will be able to • Discuss definitions of Machine Learning • Describe what major categories of ML task entail: classification, regression, clustering, relationship discovery and reinforcement learning • Discuss the relationship with Data Mining • Explain the Data Mining process • Consider current and future applications of Machine Learning and Data Mining
- 3. Module Learning Outcomes Critically appraise the performance of a learning system. Summarise and evaluate relevant research into machine learning. Analyse and evaluate the main approaches to machine learning and show similarities and differences between different approaches. Develop and appraise appropriate machine learning classification, optimisation and regression problems. Implement machine learning algorithms to classification, regression and optimisation problems specifically with Big Data. Compare and contrast different machine learning algorithms.
- 4. What is Machine Learning? 1959 Samuel, 1959:– "Field of study that gives computers the ability to learn without being explicitly programmed.“ 1978 Herbert Simon, 1978:– “Machine Learning is concerned with computer programs that automatically improve their performance through experience” 1999 Witten & Frank, 1999:– “Learning is changing behaviour in a way that makes performance better in the future.” 2017 Dangeti (2017):– “The branch of computer science that utilizes past experience to learn from and use its knowledge to make future decisions.”
- 5. Machine Learning Applications Examples • Efficient approach to attain knowledge from large amounts of data e.g. Market Basket Analysis. • Develop systems that can automatically adapt and customise themselves to individual users. E.g. Personalised news or mail filter. • Ability to mimic human and replace certain monotonous tasks e.g. In the field of quality assurance, recognising defective pieces • To develop systems that are too difficult or expensive to construct manually, since these require specific knowledge turned to a specific task (For example in the field of Medicine, a laparoscopy or another diagnosis procedure)
- 7. Machine Learning techniques • Supervised learning • Unsupervised learning • Semi-supervised learning • Reinforcement learning
- 8. Supervised Learning/ Labelled dataset • Supervised learning is training the machine with the dataset that includes input and output pairs.
- 9. • Clustering is an unsupervised learning algorithm that tries to cluster data based on their similarity. • There is no outcome to be predicted, and the algorithm just tries to find patterns in the data.
- 10. Reinforcement learning • Machine learning models to make a sequence of decisions. • Learns to achieve a goal in an uncertain, potentially complex environment. • Develop a system (agent) that improves its performance based on interactions with the environment. • Games
- 11. • Classification • Decision trees, SVMs Supervised Learning • Regression • Linear Regression, Neural nets; k NN • Clustering • k-Means, EM-clustering Unsupervised Learning • Relationship Discovery • Association Rules; Bayesian nets • Reinforcement Learning Reward Based Learning • Q-Learning, SARSA
- 12. Example Is the height of parents is passed on to their children? You want to create a model to see if the height in a population increase over time. You have two datasets: the height of the parents and the height of the children. What type of machine learning would you apply to solve this problem? a) Supervised Learning b) Reinforcement Learning c) Unsupervised Learning
- 13. Example You have a robot that needs to find his way out of a maze, it does not know the maze in advance, takes a series of decisive actions without supervision and, in the end, a reward will be given, either +1 or -1, if he finds a wall or not. Based on the final payoff/reward, the robot re evaluates its paths, and finally gets out of the maze. What type of machine learning would you apply to solve this problem? a) Supervised Learning b) Reinforcement Learning c) Unsupervised Learning
- 14. Example You have to discover groups of students based on their interests? Teachers will use this information to develop distinct engaging teaching activities What type of machine learning would you apply to solve this problem? a) Supervised Learning b) Reinforcement Learning c) Unsupervised Learning
- 15. Supervised Learning/Unsupervised Learning • Classification • Regression--> Y=Mx+C • Clustering
- 18. • Google CoLab
- 23. Output
- 28. Thank you Next Class : Statistical