CARE: Customer Automated Recommendation Engine

CARE：客户自动推荐引擎

• 批准号: 700643
• 金额: $ 3.02万
• 依托单位: WARWICK ANALYTICAL SOFTWARE LIMITED
• 依托单位国家: 英国
• 项目类别: GRD Proof of Market
• 财政年份: 2016
• 资助国家: 英国
• 起止时间: 2016 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=700643
• 关键词: CARE; Customer; Automated; Recommendation; Engine

The boom in big data and predictive analytics is well documented. The vast majority of dataare unstructured (at least 80%) and this is only set to grow both in volume and proportion.Text analytics is a growing market with sophisticated tools from well-established vendors.However these tools are either focused around specific, pre-defined tasks (e.g. sentimentanalysis) or they require skilled data scientists to use them to build models for specific tasksand then to interpret them for business users to action. This creates a huge gulf between theoceans of customer text data from call centres, CRM, POS, emails, social media etc. and theability of the companies to act on the insight in the data in real-time, both towards specificcustomers (e.g. to stop them leaving a telecoms provider) and segments of customers (e.g.prioritising actions which increase customer satisfaction, product improvement, or upsellopportunities). "CARE" or Customer Automated Recommendation Engine is a revolutionaryapproach to text analytics which does not pre-define patterns, but uses a proprietary techniqueto find any useful pattern in text and recommends the most useful predictive actions in realtime.

大数据和预测分析的繁荣是有据可查的。绝大多数数据是非结构化的(至少80%)，而且这一比例只会在数量和比例上都有所增长。文本分析是一个不断增长的市场，拥有来自知名供应商的复杂工具。然而，这些工具要么专注于特定的、预定义的任务(例如情感分析)，要么需要熟练的数据科学家使用它们为特定任务建立模型，然后解释它们，以便业务用户采取行动。这在来自呼叫中心、CRM、POS、电子邮件、社交媒体等的海量客户文本数据和公司对数据中的实时洞察采取行动的能力之间产生了巨大的鸿沟，既针对特定客户(例如，阻止他们离开电信提供商)，也针对细分客户(例如，确定提高客户满意度、产品改进或追加销售机会的优先行动)。“CARE”或客户自动推荐引擎是一种革命性的文本分析方法，它不预先定义模式，而是使用专有技术在文本中找到任何有用的模式，并实时推荐最有用的预测性操作。