Creating portfolio risk and return models [37]

Go to: Summary | Previous | Next   
Bullet points include: Within same underlying framework might add: Heteroscedasticity, autoregressive heteroscedasticity, generalised least squares and/or non-Normal innovations All still have well defined spectra akin to ARMA/ARIMA models (peaks /troughs corresponding to zeros or poles of polynomial referred to in previous slide) To get wider range of behaviours must include chaotic behaviour (adopting mathematical definition of chaos) C.f. Neural networks and other artificial Intelligence techniques Needs view about what parts of the past or what types of deductive reasoning from available observations are ‘most’ relevant to present

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile