ASTK18318U Machine learning in Conflict Prediction

Volume 2020/2021
Education

!!This Course will be 100% online!!

 

Bachelor student (2017 programme curriculum): 7.5 ECTS

Master student: 7.5 ECTS

Content

We will explore how modern computational methods from the field of machine learning can aid us in forecasting and predicting future conflicts.

 

The first part of the course will focus on giving the students substantial knowledge regarding conflict prediction. As such, we will start by surveying the literature pertaining to the current state of conflict prediction. What can we do now and what are the current challenge and limitations of computational conflict prediction. 

 

The second – and primary – part of the course will focus on giving the students the practical knowledge and technical skills needed in order to break into the field themselves. Naturally, we can only scratch the surface here but we will go through some of the most common challenges and solutions facing the novice. 

 

Thus, during the course we will cover specific topics in machine learning and data science.  Notably though, this is not an introductory course to machine learning. Point being we will only cover vary specific topics, which relate directly to the subject of conflict prediction. 

 

Troughout the course we will be working with the programming language Python along with data from the PRIO Grid and the UCDP.

Learning Outcome

Knowledge:

To understand the current opportunities, limitations and challenges and cutting edge computational conflict prediction

 

Skills:

To perform simple conflict forecasting and prediction using Python and data from the PRIO grid and the UCDP

 

Competences:

To evaluate, interpret and convey the results obtained from their own and others predictions – here with a focus on transparency regarding both strengths and limitations of said predictions

Pensum (Simply in alphabetic order - Ca. 1110 pages)

 

  • Cederman, L.-E. and Weidmann, N. B. (2017). Predicting armed conict: Time to adjust our expectations? Science, 355(6324):474 - 476.
  • Chadefaux, T. (2014). Early warning signals for war in the news. Journal of Peace Research, 51(1):5 - 18.
  • Chadefaux, T. (2017). Conict forecasting and its limits. Data Science, 1(1-2):7 - 17.
  • Chen, T. and Guestrin, C. (2016). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining, pages 785-794. ACM.
  • Colaresi, Michael, and Zuhaib Mahmood. "Do the robot: Lessons from machine learning to improve conflict forecasting." Journal of Peace Research 54.2 (2017): 193-214.
  • Croicu, M. and Sundberg, R. (2017). Ucdp ged codebook version 18.1. Department of Peace and Conflict Research, Uppsala University. (39 pages)
  • Friedman, J., Hastie, T., and Tibshirani, R. (2001). The elements of statistical learning. Springer series in statistics New York, NY, USA. Chapter 1,2,3,4,7,9,10,15 and 16. (419 pages)
  • Goldstone, J. A., Bates, R. H., Epstein, D. L., Gurr, T. R., Lustik, M. B., Marshall, M. G.,Ulfelder, J., and Woodward, M. (2010). A global model for forecasting political instability. American Journal of Political Science, 54(1):190 - 208.
  • He, H. and Garcia, E. A. (2008). Learning from imbalanced data. IEEE Transactions on Knowledge & Data Engineering, 9:1263 - 1284.
  • Hegre, H., Nyg_ard, H. M., Karlsen, J., Strand, H., and Urdal, H. (2013). Predicting Armed Conflict, 2010{2050. International Studies Quarterly, 57(2):250 - 270. 
  • Hegre, Håvard, et al. "Introduction: Forecasting in peace research." (2017): 113-124.
  • Hegre, Håvard, Håvard Mokleiv Nygård, and Ranveig Flaten Ræder. "Evaluating the scope and intensity of the conflict trap: A dynamic simulation approach." Journal of Peace Research 54.2 (2017): 243-261.
  • Hegre, Håvard, et al. "ViEWS: a political violence early-warning system." Journal of peace research 56.2 (2019): 155-174.
  • King, G. and Zeng, L. (2001a). Explaining rare events in international relations. International Organization, 55(3):693 - 715.
  • King, G. and Zeng, L. (2001b). Improving forecasts of state failure. World Politics, 53(4):623 - 658.
  • Mueller, H. F. and Rauh, C. (2016). Reading between the lines: Prediction of political violence using newspaper text. American Political Science Review, 2(112):358-375.
  • Perry, C. (2013). Machine learning and conict prediction: a use case. Stability: International Journal of Security & Development, 56(2(3)). (18 pages)
  • Schrodt, P. A. (2014). Seven deadly sins of contemporary quantitative political analysis. Journal of Peace Research, 51(2):287-300.
  • Sundberg, R. and Melander, E. (2013). Introducing the ucdp georeferenced event dataset. Journal of Peace Research, 50(4):523 - 532.
  • Tollefsen, A. F., Strand, H., and Buhaug, H. (2012). Prio-grid: A uni_ed spatial data structure. Journal of Peace Research, 49(2):363 - 374.
  • VanderPlas, Jake. Python data science handbook: Essential tools for working with data. " O'Reilly Media, Inc.", 2016. Capter 1-4, and selections from 5 (ca 400 pages)
  • Ward, M. D., Greenhill, B. D., and Bakke, K. M. (2010). The perils of policy by p-value: Predicting civil conicts. Journal of Peace Research, 47(4):363 - 375.
  • Ward, Michael D., and Andreas Beger. "Lessons from near real-time forecasting of irregular leadership changes." Journal of Peace Research 54.2 (2017): 141-156.
  • Weidmann, N. B. and Ward, M. D. (2010). Predicting conict in space and time. Journal of Conflict Resolution, 54(6):883 - 901.
  • Weidmann, Nils B., and Sebastian Schutte. "Using night light emissions for the prediction of local wealth." Journal of Peace Research 54.2 (2017): 125-140.
You should at least have taken method courses equivalent to those taught at the bachelor in Political Science, UCHP (in regards to quantitative methods that is). Even better if you have taken more advanced quantitative courses involving the subject of Data science and/or Machine Learning.

Experience with Python is good; the same applies to experience in R or similar programming languages. If you have no experience with programming, you can become acquainted with the field before the course starts – for example through online learning portals such as Datacamp, Edx, Coursera or similar. Before the course starts, I will write an email with recommended courses/exercises from the mentioned portals.

I will strive to have this mail out before the summer vacation, so you have stuff to do in between semesters ^_^

Expect significant workload (and hopefully corresponding payoff).
A combination of lectures and exercises – with a heighten focus on the practical skills and knowledge obtained through the exercises.
  • Category
  • Hours
  • Class Instruction
  • 28
  • Total
  • 28
Oral
Collective
Continuous feedback during the course of the semester
Peer feedback (Students give each other feedback)
Credit
7,5 ECTS
Type of assessment
Portfolio
Portfolio exam
Marking scale
7-point grading scale
Censorship form
No external censorship
Re-exam

- For the semester in which the course takes place: Free written assignment

- For the following semesters: Free written assignment

Criteria for exam assesment
  • Grade 12 is given for an outstanding performance: the student lives up to the course's goal description in an independent and convincing manner with no or few and minor shortcomings
  • Grade 7 is given for a good performance: the student is confidently able to live up to the goal description, albeit with several shortcomings
  • Grade 02 is given for an adequate performance: the minimum acceptable performance in which the student is only able to live up to the goal description in an insecure and incomplete manner