Optimisasi Portofolio Menggunakan Hidden Markov Model

Abstract

Portfolio selection problem (or asset allocation) in brief is to determine the proportion of capital to be invested in the se of available assets. Any proportion of capital to be invested will generate return at the end of the investment period. A decision is needed on the proportion of capital that must be invested in each as set such that, at the end of investment period return obtained as high as possible. Return obtained will very according to the stock price in each period. For that we need model that maximizes return obtained. The model used is a model of hidden Markov process which can not be observed (hidden), can be observed through a process that can be observed, with implementation measures (1) Calculated observation probability bay can forward and backward algorithm, (2) Determine hidden state by Viterbi algorithm, (3) Estimate HMM parameters by Baum-Welch algorithm. The result obtained can be used to predict changes in stock prices that resulted in changes in the obtained of return.