基于改进指纹聚类的WLAN定位优化方法

将K-means聚类算法应用到无线局域网（WLAN）位置指纹定位中，虽然可以缩短定位时间，但是容易降低定位精度。为了解决此问题，提出了基于改进指纹聚类的WLAN定位优化方法。首先根据接收信号强度标准差来优化初始聚类中心的选取,然后对指纹数据进行聚类处理,最后进行在线定位。实验结果表明，与传统的WLAN位置指纹定位方法和K-means聚类定位方法相比，基于改进指纹聚类的定位优化方法不仅缩短了定位时间，还能有效提高定位精度。     

基于两阶段法的多车型卷烟配送路径优化研究

在烟草商业企业中,物流是其重要的环节,是连接客户与烟草企业间的纽带,其卷烟配送线路设计的合理与否直接影响到卷烟的配送效率和配送成本,就如何合理的规划线路成为卷烟配送的关键问题。文章以Q烟草公司为例,运用先分组后路线的两阶段法对问题进行求解,先利用K-means聚类算法对区域进行划分,再考虑工作量均衡的条件下引入遗传算法对区域进一步调整,最后利用混合遗传算法对各配送区域进行线路优化,通过实例从线路数、车辆数、空载率、配送里程及总成本等指标上可以看出两阶段法的配送效果更优。     

基于IK-MD-SA聚类算法的电力数据审计疑点研究

为解决电力行业海量非结构化数据导致审计疑点数据效率、准确性低的难题,本文提出了一种基于迭代IK-MD-SA聚类电力大数据审计疑点算法。首先运用相异性度量算法通过构造相异性矩阵和计算均值相异性改进K-means聚类算法选择初始聚类中心,并将簇均值替换为簇中位数完成后续聚类中心迭代,以消离群点影响聚类结果的准确性。然后利用改进的蜂群算法对聚类结果进行优化,使其保证高运行效率的前提下聚类结果仍具有较高准确性。最后,通过离散性电力数据进行识别潜在疑点试验,验证了所提算法的可行性和有效性。     

论数字挖掘技术在图书馆借阅系统中的应用

本文论述了数据挖掘技术在图书馆借阅系统的应用,通过对原有管理系统现有数据的整理统计,对满足要求的数据使用K-means算法进行分割,由分类后的数据可以合理地调整不同读者群的流通图书借阅方法,优化图书资源的使用状况。     

Do Airbnb’s “Superhosts” deserve the badge? An empirical study from China

The “Superhost” badge is that Airbnb entitles the host provides good services. This study verifies Airbnb’s “Superhost” mechanism by applying text mining technologies, combined with Long Short-Term Memory (LSTM) and K-Means, to the entire dataset of tourists’ online reviews of Hangzhou city, China. Six kinds of hosts’ good services are identified, including “Three Meals or Night Snacks,” “Fruits, Drinks or Snacks,” “Travel Guides,” “Free Shuttle or Helping with Luggage,” “Chats,” and “Replies or Communications.” The study reveals the minority of “Superhosts” are mentioned of providing the majority of six kinds of good services, which means “Superhosts” do deserve the badge.     

基于K-means算法的企业服务创新方向选择研究

企业在进行服务创新过程中,首先需要考虑到企业发展战略与其自身优势能力的匹配性,结合这两点来确定服务创新的发展方向。本文研究了企业基于自身已有的优势能力,运用K-means算法建立了企业服务创新战略方向选择模型,识别并确定了企业服务创新战略方向,由此得出企业进行服务创新时的最佳服务定位策略。     

数据挖掘方法在政治风险评估中的应用研究

2008年金融危机过后,中国企业将很有可能进入另一波海外收购的浪潮。然而以往的失败案例为中国企业开展海外投资敲响了警钟。基于此背景,应用K-means聚类方法,提取国家透明组织2007年CPI（Corruption Perceptions Index）数据和世界银行2007年＂良好政府＂（good government）数据,运用两次聚类方法,得出了50个存在高政治风险的国家列表。同时采用2007年的FDI对该列表进行了进一步的检验。此列表可供＂走出去＂的中国企业作为参考。     

Segmenting the College Educated Generation Y Health and Wellness Traveler

Generation Y is considered to be the next largest cohort group to significantly impact the health and wellness travel industry. This study sought to create a profile of the Gen Y health and wellness traveler using push and pull motivations and perceived wellness. K-means clustering revealed five distinct traveler segments. Wellness was found to be a motivating factor as well as the other internal and external motivators used in this study. Men were found to travel more for escape and women for the nightlife. Opportunities to both seek and avoid contact with others were revealed. Suggestions for future marketing messages for the Gen Y wellness traveler are offered.     

A practical approach to determining critical macroeconomic factors in air-traffic volume based on K-means clustering and decision-tree classification

A given region's volume of air passengers and cargo is frequently taken to represent its economic development. This research proposes a practical methodology for investigating the inherent patterns of the relationships between air-traffic volume and macroeconomic development, utilizing data-mining techniques, including K-means clustering and Decision Tree C5.0 classification. Using the case of Taiwan from 2001 to 2014, 32 potential macroeconomic factors ascertained from a literature review were combined with air-traffic volume data to establish a 168-month dataset. After this dataset was grouped into five clusters, decision trees were implemented to determine its critical macroeconomic characteristics. The resulting four critical factors and their thresholds were the Information and Electronics Industrial Production Index (IE Index), at 83.22; National Income Per Capita, at US$3,222; Employed Population, at 10.134 million; and the Japanese Nikkei 225 Stock Average, at 10564.44. Among these, the IE Index was found to be the first critical factor relating to air-traffic volume as well as the only characteristic to distinguish Cluster V – 58 consecutive months from March 2010 to December 2014 inclusive – among others, and the reasonableness of this finding was confirmed via examination of detailed air-traffic statistics. Besides, the effectiveness of the four identified critical factors as predictive variables were validated by comparing forecasted results with actual air traffic volume from 2015 to 2016. Understanding these four critical factors and their relative importance is of great value to policymakers seeking to allocate limited resources optimally and objectively. Therefore, as an effective and efficient means of capturing significant and explainable macroeconomic factors influencing air-traffic volume, the proposed methodology can be applied to strategy formulation, operations management, and investment planning by governments, airports, airlines, and related entities.     

Claims frequency modeling using telematics car driving data

We investigate the predictive power of covariates extracted from telematics car driving data using the speed-acceleration heatmaps of Gao, G. & Wüthrich, M. V. [(2017). Feature extraction from telematics car driving heatmaps. SSRN ID: 3070069]. These telematics covariates include K-means classification, principal components, and bottleneck activations from a bottleneck neural network. In the conducted case study it turns out that the first principal component and the bottleneck activations give a better out-of-sample prediction for claims frequencies than other traditional pricing factors such as driver's age. Based on these numerical examples we recommend the use of these telematics covariates for car insurance pricing.