文章基本信息

标题：An Empirical Contextual Validation of the CapitalCubeTM Market Trading Variables as Reflected in a 10-year Panel of the S&P500: Vetting for Inference Testing
本地全文：下载
作者：Edward J. Lusk ; Edward J. Lusk ; Michael Halperin 等
期刊名称：Accounting and Finance Research
印刷版ISSN：1927-5986
电子版ISSN：1927-5994
出版年度：2016
卷号：5
期号：1
DOI：10.5430/afr.v5n1p15
语种：English
出版社：Sciedu Press
摘要：Introduction: We have accepted an Assurance engagement to examine the CapitalCubeTM market analytics platform as the lead navigation tool offered by AnalytixInsightTM. This first vetting is to examine the variables which constitute the measures used in creating decision-making inferential information. Study Precise: Our reasoning in examining the reasonability of the CapitalCube variable set is that in the Big Data world spurious associations, the bane of relevance, are expected. We examined independently the four Context Variables and the four Decision-making Variables offered by AnalytixInsight. For the former, we used Spearman r screens to eliminate firms that were dynamically not in-sync with expectations of the CapitalCube Panel as expressed through the S&P500 Panel. Further, we examined inferential power issues for extended analyses. For the Decision-making variables, we used Harman Factor results to test various a priori hypothesized profiles. Results: For the Spearman screens, we eliminated only 1% of the firms in the Panel; for the Power screens, we eliminated eight firms resulting in a Panel of 487 firms. For the Decision-making variables, the Factor profiles were strongly supportive of expectations. Impact: These eight CapitalCube variables are arguably in-sync with the empirical trajectory of the S&P500 Panel over the 10-year accrual period starting in 2002. Therefore, these CapitalCube-variables seem capable of market discrimination. This variable vetting is the critical first step in evaluating any analytic platform in the trading market milieu where Big Data rules of engagement must be serviced.
关键词：Trading Market Platform Analytics;Big Data