K-fold cross validation performance comparisons of six naive portfolio selection rules: how naive can you be and still have successful out-of-sample portfolio performance?

The so-called Benford's laws are of frequent use to detect anomalies and regularities in data sets, particularly in election results and financial statements. However, primary financial market indices have not been much studied, if studied at all, within such a perspective.This paper presents features in the distributions of S&P500 daily closing values and the corresponding daily log-returns over a long time interval, [03/01/1950 -22/08/2014], amounting to 16265 data points. We address the frequencies of the first, second, and first two significant digits and explore the conformance to Benford's laws of these distributions at five different (equal size) levels of disaggregation. The log-returns are studied for either positive or negative cases. The results for the S&P500 daily closing values are showing a remarkable lack of conformity, whatever the different levels of disaggregation. The causes of this non-conformity are discussed, pointing to the danger in taking Benford's laws for granted in large databases, whence drawing ''definite conclusions''. The agreements with Benford's laws are much better for the log-returns. Such a disparity in agreements finds an explanation in the data set itself: the index's inherent trends. To further validate this, daily returns have been simulated via the Geometric Brownian Motion and calibrating the simulations with the observed data averages and testing against Benford's laws when the log-returns distribution's standard deviation changes. One finds that the trend and the standard deviation of the distributions are relevant parameters in concluding about conformity with Benford's laws.

show abstract

“…Finance", an authoritative web site providing financial data 2 . In doing so we are in line with studies like [21,22,23,20].…”

supporting

confidence: 87%

Benford’s laws tests on S&P500 daily closing values and the corresponding daily log-returns both point to huge non-conformity

Ausloos

Ficcadenti

Dhesi

et al. 2021

Physica A: Statistical Mechanics and its Applications

View full text Add to dashboard Cite

show abstract

“…The accuracy of all the models was evaluated using an 80% training and 20% testing dataset split and fivefold cross validation [91]. Note that it is shown in [92] that a number between 5 and 10 folds usually provides similar results.…”

Section: Evaluation Of ML Methods For Supervised Classificationmentioning

confidence: 99%

Individualised Responsible Artificial Intelligence for Home-Based Rehabilitation

Vourganas

Stanković

2020

Sensors

View full text Add to dashboard Cite

Socioeconomic reasons post-COVID-19 demand unsupervised home-based rehabilitation and, specifically, artificial ambient intelligence with individualisation to support engagement and motivation. Artificial intelligence must also comply with accountability, responsibility, and transparency (ART) requirements for wider acceptability. This paper presents such a patient-centric individualised home-based rehabilitation support system. To this end, the Timed Up and Go (TUG) and Five Time Sit To Stand (FTSTS) tests evaluate daily living activity performance in the presence or development of comorbidities. We present a method for generating synthetic datasets complementing experimental observations and mitigating bias. We present an incremental hybrid machine learning algorithm combining ensemble learning and hybrid stacking using extreme gradient boosted decision trees and k-nearest neighbours to meet individualisation, interpretability, and ART design requirements while maintaining low computation footprint. The model reaches up to 100% accuracy for both FTSTS and TUG in predicting associated patient medical condition, and 100% or 83.13%, respectively, in predicting area of difficulty in the segments of the test. Our results show an improvement of 5% and 15% for FTSTS and TUG tests, respectively, over previous approaches that use intrusive means of monitoring such as cameras.

show abstract

“…The SVMR technique evaluates each run of the experiment using regression, by partitioning the data internally into training, validation, and testing components (i.e., 55% of the total data). The remaining 45% of the observed data were thereafter used for forward prediction based on the hold-out method of cross-validation (e.g., Haley, 2017). The stratified partitioning of the data using this approach ensures that each partition includes similar amount of observations from each group.…”

Section: Oceanic Hot Spots Associated With Temporal Drought Patternsmentioning

confidence: 99%

Hydrological hotspots of climatic influence in Brazil: A two-step regularization approach

Ndehedehe

Haile

Agutu

et al. 2020

Atmospheric Research

View full text Add to dashboard Cite

K-fold cross validation performance comparisons of six naive portfolio selection rules: how naive can you be and still have successful out-of-sample portfolio performance?

Cited by 16 publications

References 26 publications

Benford’s laws tests on S&P500 daily closing values and the corresponding daily log-returns both point to huge non-conformity

Benford’s laws tests on S&P500 daily closing values and the corresponding daily log-returns both point to huge non-conformity

Individualised Responsible Artificial Intelligence for Home-Based Rehabilitation

Hydrological hotspots of climatic influence in Brazil: A two-step regularization approach

Contact Info

Product

Resources

About