计算机科学 ›› 2023, Vol. 50 ›› Issue (11A): 230300029-7.doi: 10.11896/jsjkx.230300029
贾经冬, 张敏南, 赵祥, 黄坚
JIA Jingdong, ZHANG Minnan, ZHAO Xiang, HUANG Jian
摘要: 随着国家数字化建设的发展,社会治理的智能化、专业化也成为城市科技进步的基本要求,各政府系统须要对人民的诉求做到高效精确的处理。而从当前的各大政府门户网站的诉求通道收集的民众诉求信息,均是通过人工方式判断责任部门,然后将其手动分配给相关部门进行后续问题的核实和处理,大大限制了诉求处理的效率和准确性。而接诉即办智能派单算法利用人工智能和深度学习方法,基于真实的民众诉求信息数据进行训练,自动精准而高效地将诉求分派到相关部门进行后续审查处理,加快了政务处理流程的速度并大大降低了不必要的人力成本,因此该智能调度算法的研究有着重要意义。首先,通过数据去噪和脱敏,将数据进行层级拼接,构建数据标签和标准流程库以进行标签对齐。然后,基于公开数据集训练地址识别基线模型,在工单分类中提出基于类别比例采样的标签融合方法解决数据类不平衡问题,实验结果显示在基线模型的基础上提高了数十个百分点。最后,结合分类模型和地址识别模型,构建智能回复模板,完成接诉即办智能派单的全流程。
中图分类号:
[1]XIONG H,YAN H L.Research on the implementation mechanism of data driven delivery platform intelligent dispatch [J].Nankai Business Management,2022,25(2):15-25. [2]WU Z C,HONG T,SHEN Z M,et al.Research on intelligent dispatching method for communication defects under service response time constant [J].Electric Power Information and Communication Technology,2022,20(4):102-107. [3]LIU J.Research on bank intelligent dispatching system based on improved word embedding representation and convolution neural network [J].Industrial Control Computer,2020,33(4):101-104. [4]CHEN P,LI H.Research on O2O takeaway delivery route optimization based on time satisfaction [J].Chinese Journal of Management Science,2016,24(S1):170-176. [5]XU Q,XIONG J,YANG Z H,et al.Vehicle routing optimizationfor takeaway delivery based on adaptive large neighborhood search algorithm [J].Industrial Engineering and Management,2021,26(3):115-122. [6]ZHANG L Y,ZHANG J,XIAO B.Research on multi-objective O2O takeaway instant delivery route optimization considering customer priority [J].Industrial Engineering and Management,2021,26(2),196-204. [7]LI T Y,LV X N,LI F,et al.Optimization model and algorithm of delivery route considering dynamic demand [J].Control and Decision,2019,34(2):406-416. [8]OZAKAN E,WARD A R.Dynamic matching for real-time ride sharing [J].Stochastic Systems,2020,10(1):1-97. [9]BRAVERMAN A,DAI J G,LIU X,et al.Empty-car routing inridesharing systems [J].Operations Research,2019,67(5). [10]XU Z,LI Z X,GUAN Q W,et al.Large-scale order dispatch in on-demand ride-hailing platforms:a learning and planning approach [J].SIGKDD explorations,2018(Udisk):900-908. [11]AKBARPOUR M,LI S W,GHARAN S O.Thickness and information in dynamic matching markets [J].Journal of Political Economy,2019,128(3):783-815. [12]One network unified management:the ‘Shanghai plan’ for the modernization of mega-city governance[OL].http://qikan.cqvip.com/Qikan/Article/Detail?id=7103433364. [13]ZHENG Y.Smart city operating system [J].Communication of the CCF,2020(2):39-44. [14]WENG L E.Deepen the reform of ‘running at most once’ and build an integrated government service model [J].Chinese Public Administration,2019(6):2. [15]Notice of the Beijing big data work promotion on printing and distributing the Beijing smart city development action outline during the ‘Fourteenth Five-Year Plan’ period(JDDF [2001].No.1) [OL].https://www.beijing.gov.cn/zhengce/zhengcefagui/202103/t20210323_2317136.html. [16]HUANG J.Main problems and suggestions of smart city construction in Fujian Province [J].Straits Science,2022(2):95-96. [17]CUI Y M,CHE W X,LIU T,et al.Pre-training with whole word masking for Chinese BERT [J].arXiv:1906.08101,2019. [18]WANG Z N,JIANG M,GAO J L,et al.Chinese named entity recognition method based on BERT [J].Computer Science,2019,46(S2):138-142. [19]JOSHI M,CHEN D Q,LIU Y H,et al.SpanBERT:improving pre-training by-representing and predicting spans [J].Transactions of the Association for Computational Linguistics,2020(8):64-77. [20]LI X N,YAN H,QIU X P,et al.FLAT:Chinese NER using flat-lattice transformer [J].arXiv:2004.11795,2020. [21]XIONG Y J,FENG Y K,WU H,et al.Fusing label embedding into BERT:an efficient improvement for text classification[C]//Proc of Findings of the Association for Computational Linguistics:ACL-IJCNLP2021,2021.2021:1743-1750. [22]SU J L.WoBERT:word-based Chinese BERT model-ZhuiyiAI[EB/OL].(2020-09-18)[2023-09-16].https://kexue.fm/archives/7758. [23]WU X,GAO C C,LIN M,et al.Text smoothing enhance various data augmentation methods on text classification tasks [J].ar-Xiv:2202.13840,2022. [24]ROBERTSON S,ZARAGOZA H.The probabilistic relevanceframework:BM25 and beyond [J].Foundations & Trends in Information Retrieval,2009,3(4):333-389. |
|