天天干天天操天天爱-天天干天天操天天操-天天干天天操天天插-天天干天天操天天干-天天干天天操天天摸

          R語(yǔ)言機(jī)器學(xué)習(xí)學(xué)術(shù)應(yīng)用培訓(xùn)

R語(yǔ)言機(jī)器學(xué)習(xí)學(xué)術(shù)應(yīng)用
基礎(chǔ)
Theory: Features of time series data and forecasting basics

R Lab: time series objects (libraries of timeSeries, xts, & mFilters)

中級(jí)
Statistical Learning (SL):

(0.5 Hour) One-step forecasting: one-step ahead model fit

(0.5 Hour) Multi-step forecasting: recursive and direct methods

(6 Hours) Linear models: ARIMAs, ETS, BATS, GAMS, Bagged; 案例實(shí)做與寫(xiě)作范例

(5 hours) Nonlinear models: Neural Network, Smooth Transition, and AAR; 案例實(shí)做與寫(xiě)作范例

R Lab: libraries of forecast, tyDyn, vars, and MSVAR.

Research Issues: unemployment forecasting, predictability of exchange rates and asset returns.

高級(jí)
Machine Learning (ML):

(3 Hours) Tree models and SVM (Support Vector Machine)

(6 Hours) Automatic ML for forecasting time series; 案例實(shí)做與寫(xiě)作范例，涵蓋自動(dòng)化演算6個(gè)機(jī)器學(xué)習(xí)方法：

(1) DRF (This includes both the Random Forest and Extremely Randomized Trees (XRT) models.)

(2) GLM

(3) XGBoost (XGBoost GBM)

(4) GBM (gradient boost machine)

(5) DeepLearning (Fully-connected multi-layer artificial neural network, not CNN/RNN LSTM)

(6) StackedEnsemble.

(6 Hours) Econometric machine learning- Causality by ML prediction; 案例實(shí)做與寫(xiě)作范例

(3 Hours) Financial machine learning- Portfolio committees introduced; 案例實(shí)做與寫(xiě)作范例

R Lab: libraries of h2o, kera, tensorflow.

Research issues: Granger causality, volatility forecasting, portfolio selection,

economic fundamentals of exchange rates