1 / 25

An Overview of State-of-the-Art Data Modelling

An Overview of State-of-the-Art Data Modelling. Introduction. Aim. To provide researchers and practitioners with an overview of state-of-the-art techniques in data modelling. But… We will also show you how to use traditional techniques well!. Why data modelling?.

herne
Download Presentation

An Overview of State-of-the-Art Data Modelling

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. An Overview of State-of-the-Art Data Modelling Introduction

  2. Aim To provide researchers and practitioners with an overview of state-of-the-art techniques in data modelling. But… We will also show you how to use traditional techniques well! An Overview of State-of-the-Art Data Modelling

  3. Why data modelling? Increasingly important to success of many practical applications: • Engineering • Ecology • Chemistry/chemical engineering • Financial services • Crime prevention • Internet search • Systems biology • Medical diagnosis • … An Overview of State-of-the-Art Data Modelling

  4. So what is data modelling? Different things to different people. • Structuring and organising data. • Physical models of data. • Models to predict unseen data. For this course consider some examples… An Overview of State-of-the-Art Data Modelling

  5. Example 1 An Overview of State-of-the-Art Data Modelling

  6. Example 1 An Overview of State-of-the-Art Data Modelling

  7. Example 1 An Overview of State-of-the-Art Data Modelling

  8. Example 1 An Overview of State-of-the-Art Data Modelling

  9. Example 1 An Overview of State-of-the-Art Data Modelling

  10. Example 1 An Overview of State-of-the-Art Data Modelling

  11. Example 2 An Overview of State-of-the-Art Data Modelling

  12. Example 2 An Overview of State-of-the-Art Data Modelling

  13. Example 2 An Overview of State-of-the-Art Data Modelling

  14. Example 3 An Overview of State-of-the-Art Data Modelling

  15. Example 3 An Overview of State-of-the-Art Data Modelling

  16. Example 4 An Overview of State-of-the-Art Data Modelling

  17. Example 4 An Overview of State-of-the-Art Data Modelling

  18. Data modelling problems • Examples 1,2 – regression. • Example 3 – classification/pattern recognition. • Example 4 – density estimation. This course - where do you put the line? An Overview of State-of-the-Art Data Modelling

  19. Supervised vs unsupervised Do you have target data? Learning with/without a teacher Batch, incremental, sequential, online… Are all the data available initially? Are the data processed one at a time? Different types of learning An Overview of State-of-the-Art Data Modelling

  20. The course • Focus on supervised learning for regression and classification. • Cover density estimation implicitly. • Emphasis is on the concepts, ideas and tools… • …not, the detailed mathematics. An Overview of State-of-the-Art Data Modelling

  21. Day 1 • 8.30-9.00: Arrival and coffee. • 9.00-10.00: Introduction to data modelling. Curve fitting. Regression. Classification. Supervised and unsupervised learning. (Tony Dodd, Department of Automatic Control & Systems Engineering) • 10.00-11.00: Linear models. Polynomials. Radial basis functions. (Tony Dodd) • 11.00-11.30: Coffee and discussion. • 11.30-13.00: Issues in data modelling. Overfitting. Generalisation. Regularisation. Validation. Input selection. Data pre-processing. (Rob Harrison, Department of Automatic Control & Systems Engineering) • 13.00-14.00: Lunch. • 14.00-15.30: Multi-layer perceptron. (Rob Harrison) • 15.30-16.30: Coffee and discussion. An Overview of State-of-the-Art Data Modelling

  22. Day 2 • 8.30-9.00: Coffee. • 9.00-10.30: Bayesian methods. Priors. Gaussian processes. (John Paul Gosling, Department of Probability and Statistics) • 10.30-11.00: Coffee and discussion. • 11.00-12.30: MCMC methods for data modelling. (Kenneth Scerri, Department of Automatic Control & Systems Engineering) • 12.30-13.30: Lunch. • 13.30-15.00: Kernel methods. Maximum-margin classification. Support vector machines. Sparse data modelling. (Tony Dodd) • 15.00-15.30: Coffee and discussion. • 15.30-16.30: Algorithms for sequential problems. (Mahesran Niranjan, Department of Computer Science) • 16.30-17.00: Discussion and round-up. An Overview of State-of-the-Art Data Modelling

  23. Notation Inputs Input variables regression Outputs classification Targets Possible values as per y An Overview of State-of-the-Art Data Modelling

  24. Basic problem Given where e is noise. Estimate f from Density estimation requires a more complicated notation – given as required. An Overview of State-of-the-Art Data Modelling

  25. Finally… • Ask questions. • The course is for you. • Use the breaks to network and discuss your work. • Administrative matters. • Useful links http://www.shef.ac.uk/acse/research/cdmg/links/ Notes will be available at http://www.shef.ac.uk/acse/events/datamodellingcourse.html An Overview of State-of-the-Art Data Modelling

More Related