车载导航数据分析及在车险行业的应用

发布时间：2018-05-01 05:02

本文选题：车载导航 + 车险　；参考：《国防科学技术大学》2014年硕士论文

【摘要】：随着信息科学技术的不断发展,越来越多的设备可以产生数据,而硬件存储设备却越来越便宜,我们因此步入了数据爆炸式增长的时代。大数据如雨后春笋股地出现在各行各业中,许多行业已经开始对大数据进行分析,并从分析中得到了惊人的价值,如互联网行业、大型零售超市,等等。然而有很大一部分行业,他们迎来了大数据,却只是简单的存储了数据,还并没有开展对数据价值的探索。如果能够有效地使用大数据,无疑将扩大企业的竞争优势。如果一个企业忽略了大数据,并将导致在竞争中逐渐落后。伴随着我国汽车市场的飞速发展,车载导航软件近几年的发展非常迅速,使用车载导航软件行车的人越来越多,车载导航迎来了大数据。大数据是机遇,同时也是挑战,如何从导航数据中获取价值成了车载导航软件公司的难题。而与此同时,中国车险市场随着我国汽车市场发展不断扩大,竞争也越来越激烈。车险行业的竞争主要是服务与价格的竞争,归根到底是风险评估能力的竞争,而目前的广泛采用的车险定价策略存在难以区分投保人真实风险的不足。本文结合车险领域风险评估的情况以及车载导航数据的特点,提出通过对车载导航数据的分析,对用户的统计驾驶情况进行评估,将评估的结果称作为驾驶统计安全系数(简称DSCF)。该系数综合考虑了用户驾驶路程、驾驶速度、驾驶区域、夜间驾驶、疲劳驾驶等情况,是用户驾驶行为和习惯的真实体现。相对我国目前保险公司所采用的车险费率因子来说,该系数更接近用户的真实驾驶风险。车险公司可以将该系数作为保费定价的主要费率因子或者将其作为保费调整的次要费率因子,还可以将该系数与传统的费率因子相结合,对车险服务品种以及定价策略进行改进和创新。本文首先提出了驾驶统计安全系数(DSCF)的概念,然后对DSCF分析方法进行设计,主要包括驾驶统计安全评价模型和导航数据分析处理两个方面。其中驾驶统计安全评价模型主要包括评价指标体系的构建以及指标权重的分配,本文借助综合评价法,结合安全驾驶的领域知识以及现有车载导航数据,设计了评估模型的指标体系,然后采用层次分析法对指标的权重进行了分配。导航数据分析处理包括对源数据进行理解、数据选择、重组、驾驶统计分析等等。最后本文以某车载导航软件公司提供的真实数据为例,对23752个用户的DSCF进行了计算,验证DSCF分析方法的有效性。
[Abstract]:With the continuous development of information science and technology, more and more devices can produce data, while hardware storage devices are becoming cheaper and cheaper. Therefore, we have entered the era of explosive growth of data. Big data has sprung up in a variety of industries, many industries have begun to analyze big data, and from the analysis to get amazing value, such as the Internet industry, large retail supermarkets, and so on. However, a large part of the industry, they welcomed big data, but simply stored data, and has not yet begun to explore the value of the data. If can use big data effectively, will expand the competitive advantage of enterprise undoubtedly. If an enterprise neglects big data, and will lead to gradually fall behind in the competition. With the rapid development of automobile market in China, vehicle navigation software has been developing very rapidly in recent years. More and more people use vehicle navigation software to drive cars, and vehicle navigation ushered in big data. Big data is both an opportunity and a challenge. How to gain value from navigation data has become a problem for vehicle navigation software companies. At the same time, with the development of China's auto market, the competition is becoming more and more fierce. The competition in auto insurance industry is mainly the competition between service and price, and in the final analysis, it is the competition of risk assessment ability. However, the widely used auto insurance pricing strategy is difficult to distinguish the real risks of policy holders. Based on the risk assessment in vehicle insurance field and the characteristics of vehicle navigation data, this paper proposes to evaluate the statistical driving situation of users through the analysis of vehicle navigation data. The results of the assessment are referred to as the driving statistical safety factor (DSCF for short). The coefficient takes into account the driving distance, driving speed, driving area, night driving, fatigue driving and so on, which is the true embodiment of the user's driving behavior and habits. Compared with the car insurance rate factor adopted by the insurance companies in China, the coefficient is closer to the real driving risk of the users. The vehicle insurance company can use the coefficient as the main rate factor for premium pricing or as the secondary premium factor for premium adjustment, and can also combine the coefficient with the traditional rate factor. Improve and innovate the vehicle insurance service and pricing strategy. In this paper, the concept of driving statistical safety factor (DSCF) is proposed, and then the DSCF analysis method is designed, which includes two aspects: driving statistical safety evaluation model and navigation data analysis and processing. The statistical safety evaluation model of driving mainly includes the construction of evaluation index system and the distribution of index weight. This paper combines the domain knowledge of safe driving and the existing vehicle navigation data with the aid of comprehensive evaluation method. The index system of the evaluation model is designed, and then the weight of the index is allocated by the analytic hierarchy process (AHP). Navigation data analysis and processing include understanding of source data, data selection, reorganization, driving statistical analysis, and so on. Finally, taking the real data provided by a vehicle navigation software company as an example, the DSCF of 23752 users is calculated to verify the validity of the DSCF analysis method.
【学位授予单位】：国防科学技术大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：F842.634;TP311.13

【相似文献】