| 基于数据流聚类的智能交通出行信息实时推荐方法 |
 点此下载全文 |
| 引用本文：郑毅.基于数据流聚类的智能交通出行信息实时推荐方法[J].计算技术与自动化,2026,(2):164-170 |
| 摘要点击次数: 39 |
| 全文下载次数: 0 |
|
|
| 中文摘要:为准确推荐交通出行实时信息，减少交通拥堵，提高出行效率，优化出行体验，提升城市交通的智能化水平，提出了基于数据流聚类的智能交通出行信息实时推荐方法。利用孤立森林算法去除原始交通出行数据中的异常值，得到过滤后的数据流集合；引入出行用户行为权重理念，充分考虑用户出行活跃度，计算交通出行用户权重值，优先选取权重较大的用户出行数据作为Canopy算法的聚类中心，完成对Canopy算法的改进。利用改进Canopy算法对过滤后的数据流集合实施聚类分析，得到具有不同交通出行特征的数据子集；构建循环神经网络模型，利用BPTT算法对模型展开训练，将获取的具有不同特征的交通出行数据子集输入至训练好的循环神经网络模型中，其输出结果即为针对当前输入数据所生成的出行推荐信息。实验表明：该方法可以为用户实时、准确地推荐交通出行信息，帮助用户避开拥堵路段，提高出行效率。 |
| 中文关键词:数据流 智能交通 出行信息 实时推荐 Canopy算法 循环神经网络 |
| |
| Real-time Recommendation Method of Intelligent Transportation Travel Information Based on Data Stream Clustering |
|
|
| Abstract:To accurately recommend real-time transportation information, reduce traffic congestion, improve travel efficiency, optimize travel experience, and enhance the intelligence level of urban transportation, a real-time intelligent transportation information recommendation method based on data stream clustering is proposed. Using the isolation forest algorithm to remove outliers from the original transportation data and obtain a filtered set of data streams. Introducing the concept of weighted travel user behavior, fully considering user travel activity, calculating the weight values of transportation travel users, prioritizing the selection of user travel data with higher weights as the clustering center of the Canopy algorithm, and completing the improvement of the Canopy algorithm. Using the improved Canopy algorithm to perform clustering analysis on the filtered data stream set, obtaining subsets of data with different transportation characteristics. Build a recurrent neural network model, train the model using the BPTT algorithm, and input a subset of traffic travel data with different features into the trained recurrent neural network model. The output result is the travel recommendation information generated for the current input data. The experiment shows that this method can provide real-time and accurate traffic information recommendations for users, helping them avoid congested road sections and improve travel efficiency. |
| keywords:data flow intelligent transportation travel information real-time recommendation Canopy algorithm recurrent neural network |
| 查看全文 查看/发表评论 下载pdf阅读器 |
|
|
|
位访问者,今日访问