Time Series Forecasting

Eğitmen: Dr. H. Sait Ölmez

Eğitimin İçeriği:

1. Why do we need Time Series?
2. Basics of Time Series:
   1. Concepts of trend and seasonality
   2. Stationarity in Time Series
   3. Additive and Multiplicative Models
   4. White Noise and Random Walk
   5. Testing for stationarity
   6. How to make a Time Series stationary
3. Time Series Forecasting
   1. Approaches to Time Series forecasting
   2. Forecast quality metrics
   3. Model selection and validation strategies
   4. Prediction strategies
   5. Forecasting methods
      1. Moving averages
      2. Simple Exponential Smoothing
      3. Double (Holt’s linear method)  Exponential Smoothing
      4. Triple (Holt-Winter method)  Exponential Smoothing
   6. ARMA Processes
      1. Autoregressive (AR) processes
      2. Testing for Unit Root
      3. Moving Average (MA) processes
      4. Concepts of Autocorrelation (ACF) and Partial Autocorrelation (PACF)
      5. Identifying the process orders for ACF and PACF
      6. Integrated ARMA /ARIMA) models
      7. Seasonal ARIMA (SARIMA)
      8. SARIMA with exogenous variables (SARIMAX)
      9. Closing remarks

Note: Numerous examples for Time Series forecasting will be conducted throughout the course using Python and its related libraries (Statsmodels, FBProphet, Machine Learning implementations)

Eğitimin Hedef Kitlesi: Kurum içinde “Veri Analisti” ve/veya “Veri Bilimci” olarak çalışanlar, zaman serileri ile tahmin çalışmaları yapan analistler (quant’lar).

Eğitimin Seviyesi: Orta-İleri

Eğitim Süresi (Belirtiniz): 2 gün

Ek Bilgiler: Katılımcıların Python programlama dilini pratik seviyede kullanabildikleri,  Çıkarımsal İstatistik ve Doğrusal Regresyon konularına temel seviyede hakim oldukları varsayılmaktadır.