ObjectiveTo determine whether acrosome function scoring—including acrosomal enzyme (AE) levels and acrosome reaction (AR) results—can predict fertilization rate in vitro.MethodsWe examined the predictive value of acrosomal enzymes (AE) determined by spectrophotometry/N-α-benzoyl-dl-arginine-p-nitroanilide for fertilization rate (FR) in vitro in a retrospective cohort study of 737 infertile couples undergoing IVF therapy. Additionally, a meta-analysis was done for prospective cohort or case-control studies; the following summary measures were reported to expand upon the findings: pooled spearman correlation coefficient (Rs), standardized mean difference (SMD), sensitivity (SEN), specificity (SPE), positive likelihood ratio (PLR), negative likelihood ratio (NLR), diagnostic score (DS), diagnostic odds ratio (DOR), and area under the summary receiver operating characteristic curve (AUC).ResultsLower AE levels determined by spectrophotometry with a cut-off value of <25μIU/106 spermatozoa were predictive of total fertilization failure (TFF) with moderate SEN (88.23%) and low SPE (16.50%). On meta-analysis, a total of 44 unique articles were selected, but given the multiple techniques described there was a total of 67 total datasets extracted from these 44 articles, comprising 5356 infertile couples undergoing IVF therapy. The AE levels or induced AR% was positively correlated with FR (Rs = 0.38, SMD = 0.79; Rs = 0.40, SMD = 0.86, respectively). Lower AE levels or induced AR% was predictive of lower fertilization rate with moderate accuracy (AUC = 0.78, AUC = 0.84, respectively); this was accompanied by low SEN/moderate SPE (0.57/0.85), moderate SEN/moderate SPE (0.79/0.87), respectively. For AE assay, the diagnostic performance in Asia (Rs = 0.24, SMD = 0.50) was inferior to that in North America (Rs = 0.54, SMD = 0.81) and Europe (Rs = 0.46, SMD = 0.92). Cryopreserved spermatozoa (SMD = 0.20, P = 0.204) were inferior to fresh spermatozoa (SMD = 0.89, P < 0.001). Sperm preparation yielded inferior results as compared to no preparation; spermatozoa after swim up were weak relevant (Rs = 0.27, P = 0.044); and there was no correlation for spermatozoa after a discontinuous gradient (SMD = 1.07, P > 0.05). Lower AE levels determined by fluorometry or substrate assay were used for predicting lower FR with low sensitivity and high specificity; the spectrophotometry assay had an uncertain predictive value. For induced AR assay, the diagnostic performance in the other areas was inferior to that in Africa (Rs = 0.65, SMD = 1.86). No preparation or double preparation yielded inferior results as compared to one preparation (Rs = 0.41); discontinuous gradient (Rs = 0.17, SMD = 0.47) was inferior to swim up (Rs =0.65, SMD = 1.51). Nonphysiological triggers (SMD = 0.81) did not differ from physiological triggers (SMD = 0.95) in general; ZP (Rs = 0.63) or mannose (Rs = 0.59) was superior to other physiological or nonphysiological triggers; and there was no correlation for human follicle fluid, progesterone, cyclic adenosine 3′-5′-p...
Construction of a reliable stock portfolio remains an open issue in quantitative investment. Multiple machine learning models have been trained for stock portfolio selection, but their practical applicability remains limited due to the challenges posed by the characteristic of a low signal-to-noise ratio (SNR), the nature of time-series data, and non-independent identical distribution in financial data. Here, we transformed the stock selection task into a matching problem between a group of stocks and a stock selection target. We proposed a novel representation algorithm of stock selection target and a novel deep matching algorithm (TS-Deep-LtM). Then we proposed a deep stock profiling method to extract the optimal feature combination and trained a deep matching model based on TS-Deep-LtM algorithm for stock portfolio selection. Especially, TS-Deep-LtM algorithm was obtained by setting statistical indicators to filter and integrate three deep text matching algorithms. This parallel framework design made it good at capturing signals from time-series data and adapting to non-independent identically distributed data. Finally, we applied the proposed model to stock selection and tested long-only portfolio strategies from 2010 to 2017. We demonstrated that the risk-adjusted returns obtained by our portfolio strategies outperformed those obtained by the CSI300 index and learning-to-rank approaches during the same period.
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.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.