This paper uses two recently developed tests to identify neglected nonlinearity in the relationship between excess returns on four asset classes and several economic and financial variables. Having found some evidence of possible nonlinearity, it was then investigated whether the predictive power of these variables could be enhanced by using neural network models instead of linear regression or GARCH models.Some evidence of nonlinearity in the relationships between the explanatory variables and large stocks and corporate bonds was found. It was also found that the GARCH models are conditionally efficient with respect to neural network models, but the neural network models outperform GARCH models if financial performance measures are used. In resonance with the results reported for the tests for neglected nonlinearity, it was found that the neural network forecasts are conditionally efficient with respect to linear regression models for large stocks and corporate bonds, whereas the evidence is not statistically significant for small stocks and intermediate-term government bonds. This difference persists even when financial performance measures for individual asset classes are used for comparison.
Purpose
The plan‐class specific reference field concept could theoretically improve the calibration of radiation detectors in a beam environment much closer to clinical deliveries than existing broad beam dosimetry protocols. Due to a lack of quantitative guidelines and representative data, however, the pcsr field concept has not yet been widely implemented. This work utilizes quantitative plan complexity metrics from modulated clinical treatments in order to investigate the establishment of potential plan classes using two different clustering methodologies. The utility of these potential plan clusters is then further explored by analyzing their relevance to actual dosimetric correction factors.
Methods
Two clinical databases containing several hundred modulated plans originally delivered on two Varian linear accelerators were analyzed using 21 plan complexity metrics. In the first approach, each database’s plans were further subdivided into groups based on the anatomic site of treatment and then compared to one another using a series of nonparametric statistical tests. In the second approach, objective clustering algorithms were used to seek potential plan clusters in the multidimensional complexity‐metric space. Concurrently, beam‐ and detector‐specific dosimetric corrections for a subset of the modulated clinical plans were determined using Monte Carlo for three different ionization chambers. The distributions of the dosimetric correction factors were compared to the derived plan clusters to see which plan clusters, if any, could help predict the correction factor magnitudes. Ultimately, a simplified volume averaging metric (SVAM) is shown to be much more relevant to the total dosimetric correction factor than the established plan clusters.
Results
Plan groups based on the site of treatment did not show noticeable distinction from one another in the context of the metrics investigated. An objective clustering algorithm was able to discriminate volumetric modulated arc therapy (VMAT) plans from step‐and‐shoot intensity‐modulated radiation therapy plans with an accuracy of 90.8%, but no clusters were found to exist at any level more specific than delivery modality. Monte Carlo determined correction factors for the modulated plans ranged from 0.970 to 1.104, 0.983 to 1.027, and 0.986 to 1.009 for the A12, A1SL, and A26 chambers, respectively, and were highly variable even within the treatment modality plan clusters. The magnitudes of these correction factors were explained almost entirely by volume averaging with SVAM demonstrating positive correlation with all Monte Carlo established total correction factors.
Conclusions
Plan complexity metrics do provide some quantitative basis for the investigation of plan clusters, but an objective clustering algorithm demonstrated that quantifiable differences could only be found between VMAT and step‐and‐shoot beams delivered on the same treatment machine. The inherent variability of the Monte Carlo determined correction factors could not be explained solely by the modalit...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.