BPF_STAF Statistics for finance

Ekonomicko-správní fakulta
jaro 2025
Rozsah
2/2/0. 6 kr. Ukončení: zk.
Vyučováno kontaktně
Vyučující
prof. Ing. Štefan Lyócsa, PhD. (přednášející)
doc. Ing. Tomáš Výrost, PhD. (přednášející)
prof. Ing. Štefan Lyócsa, PhD. (cvičící)
doc. Ing. Tomáš Výrost, PhD. (cvičící)
Garance
prof. Ing. Štefan Lyócsa, PhD.
Katedra financí – Ekonomicko-správní fakulta
Kontaktní osoba: Iva Havlíčková
Dodavatelské pracoviště: Katedra financí – Ekonomicko-správní fakulta
Rozvrh
Čt 8:00–9:50 P102, kromě Čt 3. 4.
  • Rozvrh seminárních/paralelních skupin:
BPF_STAF/01: Čt 10:00–11:50 VT202, kromě Čt 3. 4., Š. Lyócsa
BPF_STAF/02: Čt 14:00–15:50 VT206, kromě Čt 3. 4., Š. Lyócsa
BPF_STAF/03: Čt 16:00–17:50 VT202, kromě Čt 3. 4., T. Výrost
BPF_STAF/04: Čt 18:00–19:50 VT202, kromě Čt 3. 4., T. Výrost
Předpoklady
The basic terms in calculus of probability and mathematical statistics.
Omezení zápisu do předmětu
Předmět je nabízen i studentům mimo mateřské obory.
Mateřské obory/plány
Cíle předmětu
The higher-level goal of the course is for the student to be able to understand time-series data and perform a forecasting analysis of a time-series data.
Výstupy z učení
After successfully completing the course, the student will be able to: - present time-series data, - describe basic time-series properties, - use an appropriate uni-variate model, - select appropriate methods and predict the time-series, - evaluate prediction, - perform fixed and random effect panel data estimation, - perform an event study with financial market data. - work in an appropriate statistical software.
Osnova
  • 1. Characteristics of time-series data - part I: - time-series data - visualization - descriptive statistics -> show event, show extremes, shaded areas - Auto-correlation of a time-series. - Partial auto-correlation. - Long-memory - Arch effects. - Testing for significance in serial-correlation and arch effects. 2. Characteristics of time-series data - part II: - stationarity - co-variance stationarity - weekly dependent time-series. - unit-root. - ADF test, ADF-GLS test, KPSS test. - Differencing and de-trending. 3. Forecasting non-stationary time-series - part I: - moving average. - forecasting framework - rolling & expanding rolling window - loss functions - mse & mae & mpe & ... custom choices. - cross-validation (tuning hyper-parameters). - exponential moving average. - Seasonality. - Cycle. - Trend. - Classical additive decomposition. - Multiplicative decomposition. 4. Forecasting non-stationary time-series - part II: - Holt's double exponential moving average. - Brown's double exponential moving average. - Holt and Winter's triple exponential moving average. - Linear regression model. 5. Forecasting non-stationary time-series - part III: - OLS estimation - summary - Linear regression model static models with deterministic components (trends, seasonality, breaks). - Linear regression model inference under time-series framework (auto-correlation & heteroscedasticity of residuals). - Robust inference. 6. Forecasting stationary time-series - part I: - AR(1) - AR(p) - recursive and direct forecasting - ARDL 7. Forecasting stationary time-series - part II: - MA(1). - MA(q). - Invertibility. - Maximum Likelihood Estimation & alternatives. - ARMA(p,q). - GARCH(p,q) model (forecasts). - Box-Jenkins classical & modern model selection. 8. Machine learning with linear models - part I: - Machine Learning principles. - Backward elimination. - Forward elimination. - Bi-directional elimination. - LASSO. - Ridge. - Elastic Net. 9. Machine learning with linear models - part II: - Feature engineering & data transformations. - Hyper-parameter tuning. - WLS models. - Complete subset linear regressions. - Unconditional combination forecasts. - Conditional combination forecasts. 10. Further topics in statistics in finance - part I: - Bootstrapping. - Bootstrapping in time-series models. - Binary outcome - logistic regression. - Marginal effects. - Complete subset logistic regression. 11. Further topics in statistics in finance - part II: - Panel data structures. - LSDV model. - Panel unit-root tests. - Fixed effect model. - Random effect model. - Estimation and robust inference. - Dynamic panel data alternatives. - Forecasting with panel data. 12. Introduction to event studies in finance: - The structure of an event study - Event study periods - Market models - Abnormal returns - Cumulative abnormal returns - Testing significance of abnormal returns
Literatura
    povinná literatura
  • JAMES, Gareth R., Daniela WITTEN, Trevor HASTIE a Robert TIBSHIRANI. An introduction to statistical learning : with applications in R. Second edition. New York: Springer, 2021, xv, 607. ISBN 9781071614174. info
  • HYNDMAN, RJ a G ATHANASOPOULOS. Forecasting: principles and practice. 2nd edition. Melbourne, Australia: OTexts, 2018. URL info
    doporučená literatura
  • WOOLDRIDGE, Jeffrey M. Introductory econometrics : a modern approach. Seventh edition. Boston: Cengage Learning, 2020, xxi, 826. ISBN 9781337558860. info
Výukové metody
Lectures with key principles, ideas and definitions. Seminars use computers and appropriate statistical software.
Metody hodnocení
The final grade is calculated by summing the three scores: 1) The midterm written during one of the seminars that leads from 0 to maximum of 30 points. 2) The exam - theoretical test that leads from 0 to a maximum of 30 points. 3) The exam - practical exercise that leads from 0 to a maximum of 40 points.
Náhradní absolvování
The student should contact the teacher who is leading his seminars. Usually the student should try to find a suitable alternative course. If that is not possible, student will need to take the midterm during the semester (online) and exam (on-site) with potentially an additional task to compensate for not attending semminars.
Vyučovací jazyk
Angličtina
Další komentáře
Studijní materiály
Předmět je vyučován každoročně.
  • Statistika zápisu (nejnovější)
  • Permalink: https://is.muni.cz/predmet/econ/jaro2025/BPF_STAF