Abstract:
One of the main concerns of investors is the evaluation of the return on investment, which is conducted using various models such as the CAPM (single-factor model), Fama-French three/five-factor models, and Roy and Shijin’s six-factor model and other models known as multi-factor models. Despite the widespread use of these models, their major drawbacks include sensitivity to unexpected changes, sudden shocks, high turbulence of price bubble, and so on. To eliminate such negatives, the multi-factor model using the penalty function method is used, in which, instead of averaging, the optimization and avoidance of the effects of abnormal changes and other factors affecting the capital market are considered. In order to evaluate stock returns, it is possible to select effective factors, to simulate and develop a model appropriate to the conditions governing the capital market in Iran. In the present study, by forming portfolios of investments and identifying and refining effective factors, the classification and estimation of the hybrid model of penalty and multi-factor (P & PCA) functions were performed based on the functional data during 2007-2017. The results of this study indicated that the extensive use of the simulation algorithm for the penalty function in the form of P & PCA estimation method improves the efficiency of multi-factor methods in stock return evaluation, and that the use of the hybrid algorithm of penalty and multi-factor functions, compared to the exclusive use of multi-factor models, brings a higher accuracy in estimating stock returns.
Machine summary:
Authors whose native language is not English are recom- Vol. 4, Issue 2, (2019) Advances in mathematical finance and applications [45] The Integration of Multi-Factor Model of Capital Asset Pricing and Penalty Function for Stock Return Evaluation mended to seek the advice of a native English speaker, if possible, before submitting their manu-scripts.
, (1) %%انتهای جدول%%] The penalty function can be defined in the form of an unconstrained penalty function model and within the framework of a data-mining model based on meta-heuristic algorithms and according to the findings of Ando and Bai [63], as presented below: [%%ابتدای جدول%%∅=+, (, ), (2), , , , , , ∈ , , , In this new model, we have:, , , , , , , Vol. 4, Issue 2, (2019), , , Advances in mathematical finance and applications[47], %%انتهای جدول%%] The Integration of Multi-Factor Model of Capital Asset Pricing and Penalty Function for Stock Return Evaluation ) VIew the image of this page) Therefore, g(Ci(x)) is a penalty function in the new model, in which δk is penalty coefficients.
In this classification, Vol. 4, Issue 2, (2019) Advances in mathematical finance and applications [49] The Integration of Multi-Factor Model of Capital Asset Pricing and Penalty Function for Stock Return Evaluation variables or effective factors are divided into categories of effective F = (F1 … FT )Tand hidden J = (J1 … JT )Tfactors.
This is the estimated stan- Vol. 4, Issue 2, (2019) Advances in mathematical finance and applications [51] The Integration of Multi-Factor Model of Capital Asset Pricing and Penalty Function for Stock Return Evaluation dard deviation of the parameter penalty.