摘要：【中文】本发明提出了一种医生技术水平评级方法，包括采集某一专业以及该专业下所有医生名称和每个医生名下所有病例的Drgs分组及权重值；对采集的所有数据进行清洗；训练Drgs权重分段模型；提取最优分段结果；计算每个医生的病例组合指数；通过聚类方法，以每个医生的病例组合指数和该医生每个权重分段下的病例数累计占比作为输入特征，将输入特征代入到聚类模型中，进行空间的划分，将同一个科室的不同医生根据技能水平预先划分为10个类别；提取预聚类结果；使用决策树进行验证，以及修正；进行分级，并输出最终10个级别结果；对每个医生的技能水平进行评价。本发明能够公平衡量同一家医院同一个科室医生技术水平，不同医院下同一个科室医生技术水平。 【EN】The invention provides a doctor technical level rating method, which comprises the steps of collecting a certain specialty, all doctor names under the specialty and Drgs groups and weight values of all cases under each doctor name; cleaning all the acquired data; training a Drgs weight segmentation model; extracting an optimal segmentation result; calculating case combination indexes of each doctor; by a clustering method, taking the case combination index of each doctor and the case number accumulated ratio of each weight subsection of the doctor as input characteristics, substituting the input characteristics into a clustering model, dividing the space, and dividing different doctors in the same department into 10 categories in advance according to the skill level; extracting a prepolymerization result; verifying and correcting by using a decision tree; grading and outputting final 10 grade results; the skill level of each physician was evaluated. The invention can fairly measure the technical level of doctors in the same department in the same hospital and the technical level of doctors in the same department in different hospitals.

摘要：【中文】本发明提出了一种医学数据统计查询系统及方法，包括WEB端、数据库；所述WEB端包括用户登录注册模块、数据统计分析模块、数据查询模块、消息模块、维度选择匹配模块；所述数据库包括数据获取模块、维度分类模块、用户权限模块、数据生成视图模块；户通过WEB端的用户登录注册模块进行注册，注册成功后，在数据库方面，用户权限模块会根据相应账号密码进行创建数据库用户，并进行相关的数据库权限控制；用户通过数据查询模块设定用户实际的数据筛选条件；进行数据筛选时，由维度选择匹配模块来定义约束条件，然后数据统计分析模块为用户生成折线图和饼图；本发明有利于提高用户对数据的查看效率，能够满足用户基本数据查询需求。 【EN】The invention provides a medical data statistical query system and a method, comprising a WEB end and a database; the WEB side comprises a user login registration module, a data statistical analysis module, a data query module, a message module and a dimension selection matching module; the database comprises a data acquisition module, a dimension classification module, a user authority module and a data generation view module; a user logs in a registration module through a user at a WEB end to register, and after the user successfully registers, the user authority module creates a database user according to a corresponding account password and performs related database authority control on the aspect of a database; a user sets actual data screening conditions of the user through a data query module; when data screening is carried out, a dimension selection matching module is used for defining constraint conditions, and then a data statistical analysis module generates a line graph and a pie graph for a user; the invention is beneficial to improving the data viewing efficiency of the user and can meet the basic data query requirement of the user.

摘要：【中文】本发明提供了一种数据和控制分离的TCP流漂移处理方法及装置，在负载均衡集群中部署了控制设备，每台负载均衡设备各自存储的TCP NAT连接转换表信息同时也存储在控制设备上，当负载均衡设备接收到发生流漂移的TCP报文时，会把TCP报文发送给控制设备或向控制设备发送该报文所属的TCP连接信息的查询请求，收到TCP报文的控制设备或收到控制设备反馈信息后的负载均衡设备会再将该报文发送给对应的负载均衡设备，该对应的负载均衡设备将该报文进行NAT转换后转发到目的服务器。采用本发明的负载均衡集群，可以高效管理整个集群的NAT连接转换表，降低带宽压力，降低集群成本。 【EN】The invention provides a TCP flow drifting processing method and a device with separated data and control, wherein control equipment is arranged in a load balancing cluster, TCP NAT (network address translation) connection conversion table information stored by each load balancing equipment is also stored on the control equipment, when the load balancing equipment receives a TCP message with flow drifting, the TCP message is sent to the control equipment or an inquiry request of TCP connection information to which the message belongs is sent to the control equipment, the control equipment receiving the TCP message or the load balancing equipment receiving feedback information of the control equipment sends the message to corresponding load balancing equipment, and the corresponding load balancing equipment conducts NAT conversion on the message and then forwards the message to a target server. By adopting the load balancing cluster, the NAT connection conversion table of the whole cluster can be efficiently managed, the bandwidth pressure is reduced, and the cluster cost is reduced.

摘要：【中文】本发明提出了一种寻找电子病历中诊断与药品相关性的方法，步骤如下：S1、从电子病历中获取多条诊断数据和收费明细数据；S2、基于机器学习算法对诊断数据中的诊断名称和收费明细数据中的药品项目进行归一化处理；S3、基于患者编码对归一化后的诊断名称和药品项目进行匹配；S4、利用TF‑IDF模型计算匹配后的诊断名称与药品项目的相关性得分；S5、对相关性得分进行逆排序，根据阈值判断与诊断名称强相关的药品项目并输出判断结果。本发明方法可以得到某个诊断下最常使用的、疗效最好的几个药品，计算结果准确且具有明确的指定性，有利于统计医院实际用药情况，为诊治过程中的药品使用提供数据基础。 【EN】The invention provides a method for searching the correlation between diagnosis and medicines in an electronic medical record, which comprises the following steps: s1, acquiring a plurality of pieces of diagnosis data and charging detail data from the electronic medical record; s2, normalizing the diagnosis name in the diagnosis data and the medicine item in the charge detail data based on a machine learning algorithm; s3, matching the normalized diagnosis names with the medicine items based on the patient codes; s4, calculating the correlation score of the matched diagnosis name and the medicine item by using the TF-IDF model; s5, the relevance scores are inversely ranked, and the drug items strongly correlated with the diagnosis name are determined according to the threshold value, and the determination result is output. The method can obtain a plurality of medicines which are most frequently used under certain diagnosis and have the best curative effect, has accurate calculation result and clear specificity, is favorable for counting the actual medication condition of a hospital, and provides a data basis for the medicine use in the diagnosis and treatment process.

摘要：【中文】本发明提出了一种基于Tengine学习框架的人脸识别方法，包括以下步骤：步骤1）使用caffe为框架训练完生成人脸识别caffe.proto格式的模型；步骤2）在主控制器上烧写系统镜像；步骤3）在系统上利用下载工具下载Tengine源码，安装依赖包，编译源码；步骤4）使用tengine去调用caffe.proto模型；步骤5）集成tengine与其余部分软件完成人脸识别方，本发明提高人脸识别的精度，可实现人脸精确识别，可用于闸机中。 【EN】The invention provides a face recognition method based on a Tengine learning framework, which comprises the following steps: step 1), using a cafe as a frame to train and generate a model of a face recognition cafe. Step 2) programming a system mirror image on the main controller; step 3) downloading the Tengine source codes on the system by using a downloading tool, installing a dependency package and compiling the source codes; step 4) using tengine to call a cafe. And step 5) integrating the tengine and other parts of software to finish a face recognition party.

摘要：【中文】本申请提供一种流量清洗方法、流量清洗系统和设备，包括：流量清洗设备对未命中TCP连接表的数据报文的合法性进行判断，并将合法的数据报文发送给被攻击的业务服务器，所以可以防止合法的TCP数据报文被丢失，进而防止合法TCP长连接的中断。 【EN】The application provides a flow cleaning method, a flow cleaning system and a device, comprising: the flow cleaning equipment judges the legality of the data message which does not hit the TCP connection table, and sends the legal data message to the attacked service server, so that the legal TCP data message can be prevented from being lost, and further the interruption of the legal TCP long connection is prevented.

摘要：【中文】本发明公开了一种具有减速功能的自由降放式救生艇，包括救生艇本体和船体，所述救生艇本体上的拉环通过钢制绳索连接有ax型热气球加热球囊，所述ax型热气球加热球囊的气球球囊赤道下方位置沿纵向设有四个竖的排气口，所述水平滑道内设有主动辊和从动辊，所述水平滑道的两侧和前端均设有绳索紧固机构，所述救生艇本体上的拉环通过固定绳索连接于绳索紧固机构。本发明ax型热气球加热球囊加热后，对救生艇本体具有提升的作用，起到减速效果，有效防止救生艇本体顺着滑道冲击水面，使得救生艇本体可平稳的顺着滑道脱离船体，滑至水面，达到逃生的作用。 【EN】The invention discloses a free-falling lifeboat with a speed reducing function, which comprises a lifeboat body and a lifeboat body, wherein a pull ring on the lifeboat body is connected with an ax-type fire balloon heating balloon through a steel rope, four vertical exhaust ports are longitudinally arranged at the position below the equator of the balloon of the ax-type fire balloon heating balloon, a driving roller and a driven roller are arranged in a horizontal slideway, rope fastening mechanisms are arranged at the two sides and the front end of the horizontal slideway, and the pull ring on the lifeboat body is connected to the rope fastening mechanisms through a fixed rope. After the ax-type hot air balloon heating balloon is heated, the life boat body is lifted, the speed reduction effect is achieved, the life boat body is effectively prevented from impacting the water surface along the slide way, the life boat body can be stably separated from the boat body along the slide way and slides to the water surface, and the escape effect is achieved.

摘要：【中文】本发明公开了一种探针调度方法、装置、设备和存储介质。其中，该方法包括：检测用作探针的目标服务器的状态，所述目标服务器为服务器机房中当前负载小于负载阈值的服务器，所述负载阈值根据所述服务器机房的网络类型确定；从服务器机房中未作探针的服务器中选取符合探针标准的服务器替换不符合状态要求的目标服务器。本发明实施例的技术方案通过对用作探针的目标服务器进行检测，当目标服务器性能下降时，及时选择新的服务器替换性能下降的目标服务器作为探针，保障了探针的网络质量检测性能，可提高检测的准确性，无需用户对探针重新进行手动配置，简化了探针的调度过程。 【EN】The invention discloses a probe scheduling method, a probe scheduling device, probe scheduling equipment and a probe scheduling storage medium. Wherein, the method comprises the following steps: detecting the state of a target server used as a probe, wherein the target server is a server with the current load smaller than a load threshold value in a server room, and the load threshold value is determined according to the network type of the server room; and selecting servers meeting the probe standard from servers which do not make probes in the server room to replace target servers which do not meet the state requirement. According to the technical scheme of the embodiment of the invention, the target server used as the probe is detected, and when the performance of the target server is reduced, a new server is selected in time to replace the target server with reduced performance as the probe, so that the network quality detection performance of the probe is ensured, the detection accuracy can be improved, the user does not need to manually configure the probe again, and the scheduling process of the probe is simplified.

摘要：【中文】一种自适应性管道机器人，属于机器人领域。包括行进机构和中心控制箱；所述行进机构包括多个循环相连的伸缩杆组，一个伸缩杆组的底部与下一个伸缩杆组的中部相连，行进机构的中心为伸缩杆组循环连接形成的空间，空间内固定有法兰A，法兰A与位于中心控制箱内的电机A相连，伸缩杆组的顶部设有电机B，电机B连接麦克纳姆轮。本发明通过多个间隔的麦克纳姆轮行走，具备跨越障碍的能力，伸缩杆组使本发明能够适应管径的变化。 【EN】A self-adaptive pipeline robot belongs to the field of robots. Comprises a traveling mechanism and a central control box; the traveling mechanism comprises a plurality of telescopic rod groups which are connected in a circulating mode, the bottom of one telescopic rod group is connected with the middle of the next telescopic rod group, the center of the traveling mechanism is a space formed by the telescopic rod groups in a circulating connection mode, a flange A is fixed in the space and connected with a motor A located in a central control box, a motor B is arranged at the top of each telescopic rod group, and the motor B is connected with a Mecanum wheel. The invention walks through a plurality of spaced Mecanum wheels, has the capability of crossing obstacles, and the telescopic rod group ensures that the invention can adapt to the change of pipe diameter.

摘要：【中文】本发明公开了一种轨道交通建设安全监控系统，包括视频图像监控模块、传感器数据采集模块、通信模块、控制模块、数据存储模块、异常报警模块、施工人员佩戴终端、数据显示模块和供电模块；所述视频图像监控模块与所述控制模块连接，用于采集隧道施工现场的实时视频图像信息，并将所述实时视频图像信息发送至所述控制模块等。本发明通过视频图像监控模块、传感器数据采集模块、通信模块、控制模块、数据存储模块等之间的协同配合，可以对隧道施工现场进行全自动全方位的异常监控，并对监控到的异常进行分类处理，非事故类异常采用巡检机器人处理的方式，对于事故类异常才通知施工人员前往处理，这种模式直接进一步解放的人力。 【EN】The invention discloses a rail transit construction safety monitoring system which comprises a video image monitoring module, a sensor data acquisition module, a communication module, a control module, a data storage module, an abnormity alarm module, a constructor wearing terminal, a data display module and a power supply module, wherein the video image monitoring module is connected with the data display module; the video image monitoring module is connected with the control module and used for acquiring real-time video image information of a tunnel construction site and sending the real-time video image information to the control module and the like. According to the invention, through the cooperative cooperation among the video image monitoring module, the sensor data acquisition module, the communication module, the control module, the data storage module and the like, full-automatic all-around abnormity monitoring can be carried out on a tunnel construction site, the monitored abnormity is classified and processed, the non-accident abnormity adopts a mode of processing by using a polling robot, a constructor is informed of the accident abnormity to go to process, and the mode directly relieves manpower.