“DynamicSocialNetworkFraudDetection”(动态社交网络反欺诈检测)是一个专注于动态社交网络中金融欺诈检测的图神经网络项目。该项目利用来自企业不同业务时间段的数据集,构建了一个代表用户复杂关系的全连通有向动态图。本项目的目标是通过利用图神经网络模型,包括但不限于GAT、GraphSAGE以及新型的GEARSage,分析用户间的复杂互联关系并预测欺诈活动。我们的工作通过深入理解社交网络的演化性质并应用最前沿的机器学习技术,增强了金融欺诈检测的能力。
"DynamicSocialNetworkFraud Detection" is a graph neural network project focusing on financial fraud detection in dynamic social networks. The project uses data sets from different business periods of the enterprise to build a fully connected and directed dynamic graph representing the complex relationship of users. The goal of this project is to analyze the complex interconnection between users and predict fraud activities by using graph neural network models, including but not limited to GAT, GraphSAGE and the new GEARSage. Our work has enhanced the ability to detect financial fraud through in-depth understanding of the evolutionary nature of social networks and the application of cutting-edge machine learning technologies. (2024-03-16, Python, 0KB, 下载0次)
该项目是UiPath设计的REFramework（机器人企业框架）的克隆。包括状态机、所有工作流和事务项目的克隆。我开发这个模板是为了帮助RPA开发人员使用最好的RPA框架，在python中实现您的过渡业务流程。
This project is a clone of REFramework (Robotic Enterprise Framework) designed by UiPath. Including state machines, clone of all workflows and transactional project. I developed this template to help RPA Developers to implement your transitional business process in python, using the best framework of RPA. (2024-02-27, Python, 0KB, 下载0次)
「城语」APP基于A级景区、历史古迹、文物保护单位等基础数据,利用先进的大模型能力实现智能化的Citywalk 路线规划,包括设计一条路线、生成路线攻略、生成景点的推荐理由等三大核心功能;利用大模型减少了人工编辑和推荐的工作量,并可以根据游客的需求进行个性化定制,提升了游客感知,促进文旅产业健康发展。
Based on the basic data of A-level scenic spots, historical sites, cultural relics protection units, etc., the "Citywalk" APP uses advanced big model capabilities to achieve intelligent Citywalk route planning, including three core functions: designing a route, generating a route strategy, and generating a recommendation for scenic spots; The use of the large model reduces the workload of manual editing and recommendation, and can be customized according to the needs of tourists, improving tourists perception and promoting the healthy development of the cultural and tourism industry. (2024-02-21, Python, 0KB, 下载0次)
StatefulLSTMforecast,在企业场景中,对订单或者发货的预测,通常是一个超级难度的不可控事件;如果依靠利用经典概率算法的深度学习神经网络,那么特征向量猜想及数据收集工作在实际落地过程中将成为不可能;而通过采用仅仅依靠订单或者发货的时间序列数据,通过有状态的LS...
StatefulLSTM forecast, in enterprise scenarios, predicting orders or shipments is usually a super difficult and uncontrollable event; If we rely on deep learning neural networks using classical probability algorithms, then feature vector guessing and data collection work will become impossible in the actual implementation process; By using time series data that only relies on orders or shipments, and through stateful LS (2019-09-04, Python, 150KB, 下载0次)
FraudDetection,在传统贸易企业中,销售人员经常会执行一些欺诈交易以便带来个人收益,实验利用最大似然估计的神经网络方案,用于检测当前交易的欺诈嫌疑,达到95.1%准确率,在此个案上,大大优于贝叶斯神经网络的方案。In traditional tradi...
FraudDetection, in traditional trading enterprises, salespeople often perform fraudulent transactions in order to bring personal benefits. The experiment uses a maximum likelihood estimation neural network scheme to detect suspected fraud in current transactions, achieving an accuracy of 95.1%. In this case, it is significantly superior to the Bayesian neural network scheme. In traditional tradi... (2019-09-02, Python, 268KB, 下载0次)
合成孔径雷达（SAR）是一种微波设备，能够在复杂
天气状况下产生高分辨率遥感影像。SAR具有分辨率
高、穿透性好的特点，可以全天候、全天时地获取图像信
息。SAR图像变化检测是通过对同一地点不同时刻的
SAR图像进行差异分析，得到所需的地物变化信息。
Change detection for SAR images can be achieved by clustering the differences image. Since SAR image can be interfered by speckle noise, thereby the change detection effect is affected. A method based on fuzzy clustering with spatial neighborhood information is proposed in this paper to improve clustering accuracy. (2021-03-19, Python, 196KB, 下载3次)
(中文版)深度学习(Deep Learning)是由Yoshua Bengio推出的机器学习领域的教科书，以开源的形式免费在网络上提供，主要面向学习机器学习的学生和研究者，以及没有机器学习背景但希望能快速地掌握这方面知识并在他们的产品中使用的软件工程师。
Deep Learning by Yoshua Bengio launched a textbook in the field of machine Learning, in the form of open source provided for free on the Internet, mainly for Learning machine Learning of students and researchers, and the lack of machine Learning background but hope to be able to quickly grasp the knowledge and use of software engineer in their products. (2019-07-31, Python, 13628KB, 下载4次)