Cities have limited resources that must be used efficiently to maintain their smooth operation. To facilitate efficient resource allocation and management in cities, in this paper, we study one such important problem: how long does it take to resolve non-emergency 311 service requests? We present SWIFT, a Non-emergency Response prediction system based on a recently developed structured regression model, sparse Gaussian Conditional Random Fields (GCRFs), that successfully captures the dependencies between historical and future response times. Through extensive experimentation on 311 service requests from New York City (NYC), Kansas City and Baltimore over a three and a half year period between January 2015 to June 2018, we demonstrate that our trained system is able to accurately predict future response times one week in advance using just the previous two weeks data at test time. SWIFT achieves superior prediction performance across all agencies, complaint types, and locations for all the three cities when compared to linear regression, ARIMA and Seasonal ARIMA baselines (up to a factor of 2X). The trained SWIFT system requires low computational resources and data at test time, thus making it an attractive system that can be readily deployed in practice.

城市的资源有限，必须有效利用这些资源来维持城市的平稳运转。为了促进城市中有效的资源分配和管理，本文研究了一个重要的问题：解决非紧急311服务请求需要花费多长时间？我们介绍SWIFT，一种基于最新开发的结构化回归模型（稀疏高斯条件随机场（GCRF））的非紧急响应预测系统，可以成功捕获历史响应时间与未来响应时间之间的依存关系。在2015年1月至2018年6月的三年半时间内，通过对来自纽约市（NYC），堪萨斯城和巴尔的摩的311个服务请求的广泛试验，我们证明了我们训练有素的系统能够准确地预测一个星期的未来响应时间事先只使用测试时的前两周数据。与线性回归，ARIMA和季节性ARIMA基线相比，SWIFT在所有三个城市的所有代理机构，投诉类型和地点均实现了卓越的预测性能（最高2倍）。