Course in Claims Frequency Prediction for Micromobility Insurance
300-hour online course in claims frequency prediction for micromobility insurance. 6 modules covering Royal Decree 970/2020 (personal mobility vehicles), EN 17128 (lightweight personal electric vehicles) and Python (scikit-learn, XGBoost). Assigned tutor and open enrolment.
Take the controls.
Run a real engineering trade-off from this field. Move the sliders, watch the model react in real time and beat the challenge.
What you will be able to do
Prepare claims data for e-scooters and e-bikes.
Fit base frequency and severity models.
Apply machine learning models with overfitting control.
Incorporate driving telematics variables into the model.
Translate predictions into pricing and prevention measures.
Govern the model in line with artificial intelligence regulation.
Develop a predictive claims model for micromobility.
Who it's for
- Data scientist in insurance who wants to specialise in claims frequency prediction in micromobility insurance.
- Pricing actuary who wants to specialise in claims frequency prediction in micromobility insurance.
- Micromobility analyst who wants to specialise in claims frequency prediction in micromobility insurance.
- Machine learning engineer who wants to specialise in claims frequency prediction in micromobility insurance.
- Model validator who wants to specialise in claims frequency prediction in micromobility insurance.
- Loss prevention manager who wants to specialise in claims frequency prediction in micromobility insurance.
- Engineering or science graduates looking to enter micromobility.
What you'll work with
- Royal Decree 970/2020 (personal mobility vehicles)
- EN 17128 (lightweight personal electric vehicles)
- Regulation (EU) 2024/1689 (AI Act)
- ISO/IEC 23894 (AI risk management)
- ASOP No. 23 (data quality)
- Regulation (EU) 2016/679 (GDPR)
6 modules · 300 h
1.1Shared fleets and private vehicles: two different populations1.2Exposure units: trips, minutes, kilometres and vehicle-days1.3Data sources: operator, app, workshop, emergency department and police reports1.4Under-reporting of minor claims and its effect on the model1.5Vehicle variables: power, weight, wheels and brakes1.6Environment variables: cycle lane, pavement, carriageway and surface1.7User variables: experience, declared age and night-time use1.8Seasonality, weather and time of day1.9Data quality: gaps, duplicates and imprecise geolocation1.10Workshop: building a claims database
2.1Separating frequency and severity, and why it works2.2Poisson distribution and real overdispersion2.3Negative binomial and models with excess zeros2.4Severity: lognormal, gamma and heavy tails2.5Generalised linear models applied to micromobility2.6Choice of link function and exposure offset2.7Treatment of categorical variables and grouping of levels2.8Relevant interactions: environment by user experience2.9Validation: deviance, residuals and temporal stability2.10Practice: frequency GLM on fleet data
3.1Trees, forests and gradient boosting in pricing3.2Comparison with the GLM: real gain versus complexity3.3Overfitting and blocked time-series cross-validation3.4Variable importance and partial dependence3.5Interpretable models and local explanation of predictions3.6Proxy variables for protected attributes: detection and removal3.7Probability calibration and reliability curves3.8Useful metrics: Gini, lift and calibration error3.9Deployment: monitoring model drift3.10Practice: comparing GLM and boosting on the same dataset
4.1Vehicle signals: speed, acceleration, turning and braking4.2Fall and impact detection from accelerometer data4.3Pavement versus cycle lane use via geolocation4.4Building driving indices from raw signal4.5Temporal aggregation and loss of information4.6Sensor noise and false positives4.7Combining telematics with classic variables4.8Real marginal contribution of telematics to the model4.9Legal and ethical limits on the use of these signals4.10Practice: risk index per trip
5.1Pure premium, loadings and commercial premium5.2Pricing by zone, by vehicle and by use5.3Reserving models: outstanding claims and IBNR in young portfolios5.4Scenario simulation and premium adequacy5.5Adverse selection and anti-selection in open fleets5.6Using the model to redesign the fleet and parking zones5.7Prevention interventions and measurement of their effect5.8Communicating results to the operator and the regulator5.9Portfolio monitoring and periodic recalibration5.10Practice: zone tariff based on the model
6.1Model documentation and traceability of decisions6.2Independent validation and peer review6.3Explainability requirements for the supervisor6.4Fairness testing and impact analysis by group6.5Version control and reproducibility of calculations6.6Synthetic data and privacy in development environments6.7Retraining: when and according to what criteria6.8Model risk and contingency plan6.9Presenting results to a non-technical committee6.10Final project: claims model for an urban fleet
How you'll learn
100% online training on the SEIUM campus, with downloadable material, recorded lessons and an assigned technical tutor. There are no fixed timetables: students progress at their own pace within the intake, which makes it possible to combine study with work.
Each module combines theory, guided reading of the original standard and applied exercises. The working software is Python (scikit-learn, XGBoost), R (glm), SHAP, MLflow, in free, academic or evaluation versions. The case studies are based on real frequency modelling procedures using trip data, out-of-sample validation and model fairness analysis. Assessment is continuous, module by module, plus the final integrating project.
Build it end to end
Your final project follows the real engineering workflow, step by step, reviewed by your tutor.
Prepare claims data for e-scooters and e-bikes.
Fit base frequency and severity models.
Apply machine learning models with overfitting control.
Incorporate driving telematics variables into the model.
Translate predictions into pricing and prevention measures.
Where this takes you
Requirements
- No prior qualification required.
- Internet connection and a computer capable of running the programme's software.
- Sufficient Spanish to follow technical regulatory texts.
Ready to start?
Send your application today. Our admissions team replies within 72 hours with your study plan and funding options.
Frequently asked questions
The programme consists of 300 hours in total, spread over 3 to 5 months. That is an average of 10 to 15 hours per week, compatible with a full-time job.
The programme uses Python (scikit-learn, XGBoost), R (glm), SHAP and MLflow. Free, academic or evaluation versions are used, and the tutor indicates the open-source alternative where one exists.
A SEIUM in-house diploma upon passing the continuous assessment and the integrating project. It is in-house training, not an official qualification from the Ministry.
Yes. Modules passed in a course or postgraduate diploma in the same area are recognised in the advanced programme. Check the credit recognition policy when you enrol.
Related programmes
Course in Insurance Policies for School Transport Vehicles
300-hour online course in insurance policies for school transport vehicles. 6 modules covering Royal Decree 443/2001 (safety in school and minors' transport), Royal Decree 1575/1989 (Compulsory…
Ethics in Claims Management Course
Online 300 h course in ethics in claims management. 6 modules covering Law 50/1980 on Insurance Contracts, Order ECO/734/2004 (customer service and ombudsman services) and Guidewire…
Course in Insurance Policies for Digital Infrastructure Failures
Online 300-hour course in insurance policies for digital infrastructure failures. 6 modules covering Regulation (EU) 2022/2554 (DORA), Directive (EU) 2022/2555 (NIS2) and CyberCube. Assigned tutor and…
More in Insurance, Risk & Actuarial Data
All Insurance, Risk & Actuarial Data programmes → · How to apply · Fees & financing · Scholarships