385 research outputs found

    Statistical Comparisons of Multiple Classifiers

    No full text
    This paper discusses the issue of comparing multiple classifiers, applied to the same test dataset of a classification problem. Assume that the output is 0 if a classifier correctly classifies a test feature point and the output is 1 otherwise.Then all the outputs from a given classifier constitute a sample of 0 and 1, and all the samples are correlated. From these dependent samples, we use Cochran's Q statistic, as an overall test statistic, to detect whether or not the error rates of the classifiersare significantly different. When the null hypothesis that the error rates are equal is rejected, a thorough analysis of the nature of the error rates, such as the ranking of the error rates, is undertaken. For this purpose, we employ the Scheffe and Bonferroni multiple comparison procedures, based on dependent samples. We also use examples to demonstrate how to make these statistical comparisons.Cheng, Xiuzhen; Chen, Dechang. (2001). Statistical Comparisons of Multiple Classifiers. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/215473

    Evaluación económica de la iniciativa de "Control de tuberculosis en grandes ciudades" en Lima - Perú

    No full text
    Este estudio tuvo tres objetivos: primero, determinar la efectividad incremental de la intervención focalizada en comparación con el programa existente; segungo, determinar el costo de efectividad y el costo de utilidad de la intervención focalizada en comparación con el programa existente; y, tercero, determinar el costo-beneficio de la intervención focalizada en comparación con el programa existente. Utilizamos una perspectiva de proveedor. Para el primer objetivo, realizamos un análisis ex-post durante los primeros 5 meses de implementación para estimar la efectividad de las intervenciones y un análisis ex-ante para los siguientes veinte años utilizando un modelo epidemiológico compartimental para estimar el impacto de la intervención focalizada en la incidencia de tuberculosis en el horizonte de veinte años. Para el segundo objetivo, hicimos un análisis de costo-efectividad para determinar la relación costo-efectividad incremental (ICER) de la intervención focalizada para evitar nuevos casos de TB. Además, para el análisis de costo-utilidad, hemos calculado los años de vida ajustados por discapacidad (AVAD) para determinar el costo por AVAD evitado. Para el tercer objetivo, hicimos un análisis de costo-beneficio donde calculamos los beneficios en términos del costo de la enfermedad, y comparamos este costo tanto para la intervención focalizada como para el status quo, para determinar el retorno neto de la inversión.Tesi

    Set analysis of coincident errors and its applications for combining classifiers

    No full text
    Although addressed in many papers, classifier dependency is still not well defined. Continuously being described by variety of statistical models from conditional probability to diversity measures, dependency among classifier out-puts was recently shown to have a crucial impact on the performance of multi-ple classifier system. However, individual classifier performances still represent competitive and simple information clearly related to the performance of the combined system. In this work we show that all the measures related to classi-fier outputs can be reformulated to represent just different forms of the same in-formation of error coincidences. Applying set analysis for the representation and description of error coincidences we define collection of classifier sets de-composed into two complementary types of coincidence levels. Furthermore we illustrate a high flexibility of using the coincidence levels, which supported be a simple algebra cover many established dependency measures including combin-ing error in case of majority voting. Moreover we show that in the sets-collection representation of error coincidences a specific inclusion relation re-sults in a quicker and more effective handling of dependency information under different complexity conditions. In the experimental section we examine rela-tions of the introduced error coincidence levels with majority voting combiner using real datasets and classifiers and indicate further potential applications of the presented concepts

    The 1126 Ma volcanic event in the Dechang Area, SW Yangtze Block, and its significance

    No full text
    Traditionally, the strata of the Luonie Valley, Dechang County, SW Sichuan, China, are considered to contain a suite of felsic volcanic rocks (the Huili Group) that erupted after c. 1050 Ma. However, we report here new age constraints, elemental and Lu–Hf isotope geochemistry for a different suite of older basaltic agglomerate lava, basaltic tuff lava and basalt from the same area, which we name the Luonie Formation. New dating results show that the basaltic volcanic suite of the upper part of the Luonie Formation formed at 1126.1 ± 9.9 Ma, significantly earlier than deposition of the Huili Group, but comparable in age to the 1142 ± 16 Ma Laowushan Formation in central Yunnan Province. Granite intrusion into the Luonie Formation dated 1050.7 ± 12.7 Ma provides crucial supporting evidence for this earlier depositional age. We also report a maximum sedimentary age of c. 1158 Ma for the underlying arkose, implying stratigraphic conformity with the basaltic volcanic rock suite. The ϵHf(t) values of the basaltic volcanic rocks are mainly positive, indicating that the rocks are mainly derived from the depleted mantle and slightly stained by crustal materials. The characteristics of P*, Nb* and Zr* anomalies also support this view. The distribution patterns of trace and rare earth elements indicate that the basaltic volcanic rocks formed in an extensional setting. The Zr/4–Y–2Nb and Th–Nb/16–Zr/117 discrimination diagrams also provide evidence for this understanding. Lithofacies analysis shows that basaltic volcanic wrocks with the characteristics of both continental and marine facies should be formed in a littoral–neritic environment. We propose here that the evidence is consistent with a phase of continental extension that preceded the convergence of the SW Yangtze Block to form part of Rodinia

    Refining stoichiometric approaches to trace soil organic matter sources

    No full text
    In their Letter‐to‐Editor, Liu & Chen (2024) raised the importance and suitability of our recent stoichiometric approach to assess the contribution of plant residue (litter) and microbial sources to mineral‐associated organic matter (MAOM). We highly appreciate their Letter as well as the further suggestion and comments and offer two points of clarification. We call for staying aware of methodological benefits and limitations of each approach, and further develop and compare multiple methods to develop a clearer picture of where and to what extent MAOM is formed from direct vs. indirect (via the microbial funnel) plant inputs.National Natural Science Foundation of China https://doi.org/10.13039/50110000180

    Xiao ji you sheng xue lun li tan tao

    No full text
    陳德昌.Running title: 消極優生學倫理探討.Thesis (M.Div.)--香港中文大學宗敎及神學學部.Running title: Xiao ji you sheng xue lun li tan tao.Includes bibliographical references (leaves 52-61).Chen Dechang.Thesis (M.Div.)--Xianggang Zhong wen da xue zong jiao ji shen xue xue bu.Chapter 一. --- 導 言 --- p.1-6Chapter 二. --- 遺傳病理學的發展 --- p.7-12Chapter 三. --- 從分配公義的原則看消極優生的必需 --- p.13-21Chapter 四. --- 從病者的人生尊嚴看消極優生的必要 --- p.22-25Chapter 五. --- 對未來世代的責任 --- p.26-30Chapter 六. --- 消極優生手段之産前檢驗的準確性 --- p.31-36Chapter 七. --- 基因工程發展與基督教神學思想 --- p.37-50Chapter 八. --- 結論 --- p.51Chapter 九. --- 附注 --- p.52-61Chapter 十. --- 注釋 --- p.62-63Chapter 十一. --- 附表 --- p.64-6

    A postoperative tumor-specific death prediction model for patients with endometrial cancer: a retrospective study

    No full text
    Background: Endometrial cancer (EC) is an epithelial malignancy occurring in the endometrium, with a 5-year mortality rate of above 10%. However, there is currently a lack of studies exploring the potential of a predictive model of tumor-specific death after surgery in these patients. Methods: From January 2015 to December 2017, data related to 482 patients with EC admitted to the Dushu Lake Hospital Affiliated to Soochow University were analyzed. Patients were divided into death (n=62) and survival (n=420) groups according to whether tumor-specific death occurred at 5 years postoperatively or not. The clinical characteristics of the two groups were compared, and the risk factors for tumorspecific death in patients with EC 5 years after surgery were investigated by logistics regression analysis. A nomogram prediction model was established according to the relevant risk factors. Results: Tumor size, Ki-67 positive rate, Federation International of Gynecology and Obstetrics (FIGO) stage, and the rate of vascular tumor thrombus between the two groups (P<0.05) were found to be the statistically significant factors. Positive Ki-67, tumor size >3.35 cm, stage III, and vascular tumor thrombus were factors that influenced the tumor-specific death at 5 years after surgery (P<0.05). The predictive model obtained an area under the receiver operating characteristic (ROC) curves in the training and verification sets of 0.847 [95% confidence interval (CI): 0.779-0.916] and 0.886 (95% CI: 0.803-0.969), respectively. Conclusions: The nomogram prediction model, which was established in this study, was proved to be valuable in predicting tumor-specific death 5 years after the surgery in patients with EC
    corecore