The subject Analysis of Spatio-Temporal Data is an optional subject that is studies in the 4th year of the Artificial Inteligence degree. In this subject we will analyze the main characteristics of spatio-temporal data and we will study the main problem and algorithm that are used to extract useful information from this type of data.



The subject is based on the knowledge acquired in the previous Data Mining and Advanced Machine Learning subjects and will be built on top of the concepts learnt in those subjects. Additionally and transversely, programming skills and statistical basics will also be required.



Nowadays, in more and more domains, the data collected by sensors and other technologies have a spatio-temporal nature. Extracting useful information from this type of data has become a major challenge for practitioners and the scientific community and, so, in the past decades a great variety of new problems and algorithms, specific for this type of data, have been proposed. Having some knowledge on these problems and algorithms is a must for experts in Artificial Inteligence.

Analyze the main characteristics of spatio-temporal data and be acquainted with the methods fort the exploratory analysis and visualization of this type of data.



Understand the main data mining problems for spatio-temporal data and know how to identify them in real scenarios.



Know about the basics of the main algorithms for different spatio-temporal data mining problems.



Design, implement and carry out small experimentation processes using R and/or Python.



Obtain conclusions from the results or visualizations obtained and provide arguments based on evidence.

1.- Introduction to spatio-temporal data

1.1.- Main characteristics of spatio-temporal data.

1.2.- Representation and visualization of spatio-temporal data



2.- Time series forecasting

2.1.- Classical statistical models

2.2.- Models based on Machine Learning

2.3.- Applications



3.- Non-supervised classification of time series

3.1.- Distances for time series

3.2.- Applications



4.- Supervised classification of time series

4.1.- Main taxonomies to categorize the algorithms and examples.

4.2.- Extensions of the basic problem of supervised classification of time series

4.3.- Applications



5.- Analysis of spatio-temporal data

5.1.- Common problems and algorithms

5.2.- Applications

Lectures with the conceptual contents of the subject will be complemented with exercises and/or examples. Additionally, in this subject we will promote the autonomous work of the student and parts of the contents will be studies by means of group projects. For this, computer and bibliographic resources will be provided that will help the students to understand the topic. The conputational aspect of the subject will be carried out using R and/or Python.

• Final Assessment System
• Tools and qualification percentages:
  • The assessment types and conditions are indicated below: zehazten dira (%): 100

The student can be evaluated under two types of assessments: continuous or final. The continuous assessment system is prioritized, as indicated in the regulation of the UPV/EHU.



If a student who meets the requirements of continuous assessment wishes to opt for the final assessment, he or she must inform the lecturers responsible for the subject in the following manner and within the following deadlines: via email once the written test of the 2nd fixed week has been graded.





CONTINUOUS ASSESSMENT



The continuous assessment involves activities that will be carried out throughout the quarter: practical individual and group works (70%), theory/practice partial exam of to be carried out the laboratory (30%).



The final mark will be the weighted mean of the results obtained in all the evaluation items, provided that a minimum of 4 has been obtained in the exam. The subject will be passed with an average mark of 5 or more.



Not taking part in the exam or not giving in the required projects will be considered a withdraw from this type of assessment.



FINAL ASSESSMENT



The final assessment involves a theory/practice partial exam of to be carried out the laboratory (50%) and some individual project that must be given in the day of the exam (50%).



The final mark will be the weighted mean of the results obtained in all the evaluation items, provided that a minimum of 4 has been obtained in the exam. The subject will be passed with an average mark of 5 or more. Not taking part in the exam or not giving in the required projects will be considered a withdraw from this type of assessment.

The assessment in extraordinary examination will be the same as the global assessment:



This type of assessment involves a theory/practice partial exam of to be carried out the laboratory (50%) and some individual project that must be given in the day of the exam (50%).



The final mark will be the weighted mean of the results obtained in all the evaluation items, provided that a minimum of 4 has been obtained in the exam. The subject will be passed with an average mark of 5 or more. Not taking part in the exam or not giving in the required projects will be considered a withdraw from this type of assessment.

There is no special required material. The student will build his/her own material following the classes and with the material provided in eGela.

Basic bibliography

Hyndman, R.J., & Athanasopoulos, G. (2021) Forecasting: principles and practice, 3rd edition, OTexts: Melbourne, Australia. OTexts.com/fpp3.



Philippe Esling and Carlos Agon. 2012. Time-series data mining. ACM Comput. Surv. 45, 1, Article 12 (November 2012), 34 pages. https://doi.org/10.1145/2379776.2379788



Bagnall, A., Lines, J., Bostrom, A. et al. The great time series classification bake off: a review and experimental evaluation of recent algorithmic advances. Data Min Knowl Disc 31, 606–660 (2017).



Hamdi, A., Shaban, K., Erradi, A. et al. Spatiotemporal data mining: a survey on challenges and open problems. Artif Intell Rev 55, 1441–1488 (2022).

Journals

Journals in the area: IEEE Transaction on Knowledge and Data Engineering, Data Mining and Knowledge Discovery, Journal of Machine Learning Research, Pattern Recognition, Knowledge based Systems, etc.

Web addresses

https://cran.r-project.org/web/views/TimeSeries.html

https://cran.r-project.org/web/views/SpatioTemporal.html

https://www.timeseriesclassification.com/