This paper focuses on the development of high-dimensional risk models to construct portfolios of securities in the US stock exchange. Investors seek to gain the highest profits and lowest risk in capital markets. We have developed various risk models, and for each model different investment strategies are tested. Out-of-sample tests are performed on a long-term horizon from 1970 until 2023.
This paper focuses on the application of quantitative portfolio management by using integer programming and clustering techniques. Investors seek to gain the highest profits and lowest risk in capital markets. A data-oriented analysis of US stock universe is used to provide portfolio managers a device to track different Exchange Traded Funds. As an example, reconstructing of NASDAQ 100 index fund is presented.
The paper examines the performance of regression models (OLS linear regression, Ridge regression, Random Forest, and Fully-connected Neural Network) on the prediction of CMA (Conservative Minus Aggressive) factor premium and the performance of factor timing investment with them. Out-of-sample R-squared shows that more flexible models have better performance in explaining the variance in factor premium of the unseen period, and the back testing affirms that the factor timing based on more flexible models tends to over perform the ones with linear models. However, for flexible models like neural networks, the optimal weights based on their prediction tend to be unstable, which can lead to high transaction costs and market impacts. We verify that tilting down the rebalance frequency according to the historical optimal rebalancing scheme can help reduce the transaction costs.
The default prediction of credit card holders in finance industry will help the financial firms, dealing with credits and monthly cash flows, to have not only a develop a better marketing plan to target the best clients but also to form a precise risk management system. In this research, the performance of several machine learning methods are evaluated to predict the default possibility of 30,000 clients in Taiwan. To do so, at first the most suitable criteria to predict the default risk are proposed according to the literature review and past works done by the noble machine learning experts. Then, after some data analysis and dataset review, two approached are used to answer the research question. Since about 10% of the card holders didn’t default during past six months but they defaulted in the seventh payment, in this research it’s tried to see if we can assume them as non-defaulters (or not) to come with a better classifier.
In previous papers regarding market microstructure and price information in the stock market, the VAR model (vector auto regressions) has been frequently mentioned as a benchmark, and there are various discussions about the pros and cons of this model. Here, we would like to extend the usage of this model to the cryptocurrency market and see how VAR performs in a new setting. Our result shows that although VAR might yield some gaps in terms of price prediction, it still catches the trend with a fair amount of accuracy and could still serve as a potential benchmark model.
The performance evaluation and prioritisation of knowledge-based firms which are working in science and technology parks enable the decision makers (DMs) to specify the benchmarks (the best firms) and grant them or help other firms to improve their performance. In this paper, the performance of new technology-based firms (NTBFs) in Iran is evaluated through a multi-criteria decision-making method. To do so, at the first the most important criteria to assess the performance of the NTBFs are proposed according to the literature and experts' opinions. Then, due to complex relationship and correlation among these criteria, an integrated fuzzy analytic network process (ANP) is proposed to evaluate NTBFs under uncertainty of evaluation parameters. The results show that the applied approach is a suitable tool for prioritisation of NTBFs.