CAT/2: Big Data Content Analytics with particular reference to Probabilistic Neural Topic Models

CAT/2：大数据内容分析，特别参考概率神经主题模型

• 批准号: 103652
• 金额: $ 82.21万
• 依托单位: MICROPORTS.NET LIMITED
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2017
• 资助国家: 英国
• 起止时间: 2017 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=103652
• 关键词: CAT; Big; Data; Content; Analytics

There is significant need across the public & private sectors for a new, cost effective means by which unstructured data can be presented to professional users. Specifically, to more instantly and easily derive meaning over time from within the ‘Free Text Box’ (FTB) content of Line of Business (LOB) data systems. There is also a critical need for better access to ‘joined-up’ relevant & predictive intelligence across domains such as the medical, social care and education divide (e.g. early warning signs for children at risk, risk evaluation of our ageing population, early onset risk of long term debilitations like dementia & mental health). The problem is that the critical ‘case notes’ information is hidden within the the content of FTBs (unstructured data) rather than in the structured data input to and held within LOB systems. This project will unlock the hidden meaning from within FTB collections and provide predictive analyses and innovative user experiences to better support evidence-based understanding and decisions. So we will make sense and more digestible the information and its relative importance that is currently largely invisible within Clinician Notes in a Hospital PAS, GP Notes in their Practice System, Adult and Child Notes in the Council’s Care Management System, Legal Margin Notes in Case Management Systems, Governance Issues within Complaints Systems, Customer Issues within CRM systems and so on. This will greatly improve the qualitative & quantitative Informatics & Business Intelligence available to an organisation for scarce-resource risk planning, service delivery and decision-making.

公共和私营部门都非常需要一种新的、具有成本效益的方法，通过这种方法可以将非结构化数据呈现给专业用户。具体而言，随着时间的推移，更即时、更轻松地从业务线（LOB）数据系统的“自由文本框”（FTB）内容中获取意义。此外，还迫切需要更好地获得跨领域的“联合”相关和预测性情报，如医疗，社会护理和教育鸿沟（例如风险儿童的早期预警信号，我们老龄化人口的风险评估，长期衰弱的早期发病风险，如痴呆症和心理健康）。问题在于关键的“案例说明”信息隐藏在FTB（非结构化数据）的内容中，而不是在输入到LOB系统并保存在LOB系统中的结构化数据中。该项目将解锁FTB收藏中隐藏的意义，并提供预测分析和创新的用户体验，以更好地支持基于证据的理解和决策。因此，我们将使信息及其相对重要性更有意义和更容易消化，这些信息目前在医院PAS中的临床医生注释、其实践系统中的GP注释、理事会护理管理系统中的成人和儿童注释、病例管理系统中的法律的保证金注释、投诉系统中的治理问题中基本上是不可见的，CRM系统中的客户问题等。这将大大提高组织可用于稀缺资源风险规划，服务交付和决策的定性和定量信息学和商业智能。