检索结果-内蒙古大学图书馆

Detection of price bubbles in Istanbul housing market using lstm autoencoders: a district-based approach

SOFT COMPUTING 2021年第12期25卷 7957-7973页

作者： Ayan, Ebubekir Eken, Suleyman Kocaeli Univ Dept Business Adm TR-41001 Izmit Turkey Kocaeli Univ Dept Informat Syst Engn TR-41001 Izmit Turkey

Since the early 2000s, there has been a long-term price increase trend in the Istanbul housing market, and this situation also has led to price bubble speculations. Since the housing sector was caught with a high level of unsold housing stock to the economic slowdown emerging in the second half of 2018, housing price bubble speculations have increased even more, especially for the Istanbul market. In this period, housing loan interest reduction campaigns were implemented by the government through state banks to stimulate the housing demand, and a probable collapse in the housing market was prevented. On the other hand, house prices continued to rise during this period due to the stimulated demand. In this paper, we perform a price bubble research on the selected districts in the Istanbul housing market over the 2007-2019 period using lstm autoencoder model. The first analysis on monthly data is performed by using housing price index, housing rent index, consumer prices index, stock market index, return on government debt securities, USD/TRY exchange rates, BIST price index, monthly deposit interest rates, mortgage interest rates and consumer confidence index, and the second analysis on quarterly data is carried out by adding building construction cost index and GDP data to the previous dataset. In the first analysis, the bubble formations differ regionally and periodically and disappeared toward the end of 2019 in some districts, while in the second analysis, the housing bubble formations have a more common and continuous appearance. Experimental results show that lstm autoencoder model can be used to detect housing bubbles effectively.

关键词： Deep learning Housing bubble Housing market Housing pricing lstm autoencoder Price bubble

来源：评论

学校读者我要写书评

暂无评论

An Event Frequency-Inverse Session Frequency Based Weighting Mechanism for Data Embedding Methodologies

引用

SOFTWARE-PRACTICE & EXPERIENCE 2025年第6期55卷 983-1002页

作者： Tasgetiren, Nail Safak, Ilgin Aktas, Mehmet S. Hepsiburada R&D Ctr Istanbul Turkiye Fibabanka R&D Ctr Istanbul Turkiye Yildiz Tech Univ Comp Engn Dept Istanbul Turkiye

ContextWith the rapid increase in data complexity and volume, analyzing and understanding large-scale datasets has become a critical challenge, especially in domains like e-commerce. This paper addresses this challenge by introducing a novel business process that utilizes an event frequency-inverse session frequency (EF-ISF) based weighting mechanism for embedding user navigational behavior in e-commerce *** primary research problem tackled in this study is the detection of patterns and distributions within user behavior data, which is often complicated by dynamic page content and session sparsity. We aim to enhance the interpretability and accuracy of user behavior analysis by integrating EF-ISF with various embedding techniques, including Word2Vec, Node2Vec, DeepWalk, Struc2Vec, and lstm *** approach is particularly effective in datasets where page refreshes lead to event repetitions, where sequential events with temporal dependencies are prevalent, and where event distributions across sessions are sparse. By applying this methodology across multiple e-commerce datasets, we observe consistent performance in clustering quality, successfully detecting underlying patterns and groups within the *** prototype implementation demonstrates the practical utility of the proposed method, with evaluation metrics confirming that EF-ISF weighting not only enhances the embeddings but also facilitates a more nuanced and significant segmentation of users' browsing *** findings provide a valuable contribution to the field of e-commerce data analytics, offering businesses a refined approach to optimizing user experience and engagement through a deeper understanding of user behavior patterns. This methodology can significantly improve the quality of insights derived from user behavior data, ultimately leading to better decision-making and enhanced user satisfaction.

关键词： clickstream data Deepwalk E-commerce embedding frequency-based weighting lstm autoencoder Node2vec Struc2vec user navigational behavior Word2vec

来源：评论

学校读者我要写书评

暂无评论

A deep semi-supervised machine learning algorithm for detecting transportation modes based on GPS tracking data

引用

TRANSPORTATION 2024年 1-21页

作者： Sadeghian, Paria Golshan, Arman Zhao, Mia Xiaoyun Hakansson, Johan Dalarna Univ Sch Technol & Business Studies SE-79188 Falun Sweden

Transportation research has benefited from GPS tracking devices since a higher volume of data can be acquired. Trip information such as travel speed, time, and most visited locations can be easily extracted from raw GPS tracking data. However, transportation modes cannot be extracted directly and require more complex analytical processes. Common approaches for detecting travel modes heavily depend on manual labelling of trajectories with accurate trip information, which is inefficient in many aspects. This paper proposes a method of semi-supervised machine learning by using minimal labelled data. The method can accept GPS trajectory with adjustable length and extract latent information with long short-term memory (lstm) autoencoder. The method adopts a deep neural network architecture with three hidden layers to map the latent information to detect transportation mode. The proposed method is assessed by applying it to the case study where an accuracy of 93.94% can be achieved, which significantly outperforms similar studies.

关键词： Travel identification lstm autoencoder Unsupervised learning Deep learning GPS tracking data

来源：评论

学校读者我要写书评

暂无评论

Edge AI for Real-Time Anomaly Detection in Smart Homes

引用

FUTURE INTERNET 2025年第4期17卷 179-179页

作者： Reis, Manuel J. C. S. Serodio, Carlos Univ Tras Os Montes & Alto Douro Engn Dept P-5000801 Vila Real Portugal IEETA Inst Elect & Informat Engn Aveiro P-3810193 Aveiro Portugal Algoritmi Ctr P-4800058 Guimaraes Portugal

The increasing adoption of smart home technologies has intensified the demand for real-time anomaly detection to improve security, energy efficiency, and device reliability. Traditional cloud-based approaches introduce latency, privacy concerns, and network dependency, making Edge AI a compelling alternative for low-latency, on-device processing. This paper presents an Edge AI-based anomaly detection framework that combines Isolation Forest (IF) and Long Short-Term Memory autoencoder (lstm-AE) models to identify anomalies in IoT sensor data. The system is evaluated on both synthetic and real-world smart home datasets, including temperature, motion, and energy consumption signals. Experimental results show that lstm-AE achieves higher detection accuracy (up to 93.6%) and recall but requires more computational resources. In contrast, IF offers faster inference and lower power consumption, making it suitable for constrained environments. A hybrid architecture integrating both models is proposed to balance accuracy and efficiency, achieving sub-50 ms inference latency on embedded platforms such as Raspberry Pi and NVIDEA Jetson Nano. Optimization strategies such as quantization reduced lstm-AE inference time by 76% and power consumption by 35%. Adaptive learning mechanisms, including federated learning, are also explored to minimize cloud dependency and enhance data privacy. These findings demonstrate the feasibility of deploying real-time, privacy-preserving, and energy-efficient anomaly detection directly on edge devices. The proposed framework can be extended to other domains such as smart buildings and industrial IoT. Future work will investigate self-supervised learning, transformer-based detection, and deployment in real-world operational settings.

关键词： edge AI anomaly detection smart home IoT isolation forest lstm autoencoder

来源：评论

学校读者我要写书评

暂无评论

lstm-AE-WLDL: Unsupervised lstm Auto-Encoders for Leak Detection and Location in Water Distribution Networks

引用

WATER RESOURCES MANAGEMENT 2023年第2期37卷 731-746页

作者： Kammoun, Maryam Kammoun, Amina Abid, Mohamed Sfax Univ Natl Engn Sch Sfax CES Res Lab Sfax 3038 Tunisia

Basically, leaks and faults in water distribution pipelines beget fairly severe water loss and affects largely its potability. For this reason, leakage detection is extremely significant for the preservation of water resources and quality. This paper introduces a novel unsupervised RNN model for leakage detection and location. The elaborated approach relies upon a multivariate lstm autoencoder, as well as a multithresholding to monitor all water distribution network zones. A threshold for each measurement point of water distribution network is determined to identify anomaly in hydraulic data and detect leak events. Furthermore, a statistical study is conducted to estimate the leak locations' area. Both flow and pressure data from different realistic water demands scenarios of the LeakDB benchmark are assessed. Experiment results corroborate the effectiveness and reliability of the proposed system for both data types. Detection sensitivity achieved 97% using pressure data and 100 % using flow data, with true leak zone identification for 95% of scenarios.

关键词： Leak detection Leak area location Unsupervised RNN lstm autoencoder LeakDB

来源：评论

学校读者我要写书评

暂无评论

lstmAE-DWSSLM: A unified approach for imbalanced time series data classification

引用

APPLIED INTELLIGENCE 2023年第18期53卷 21077-21091页

作者： Liu, Jingjing Yao, Jiepeng Zhou, Qiao Wang, Zhongyi Huang, Lan China Agr Univ Coll Informat & Elect Engn Beijing 100083 Peoples R China Minist Agr Key Lab Agr Informat Acquisit Technol Beijing Beijing 100083 Peoples R China Minist Educ Key Lab Modern Precis Agr Syst Integrat Beijing Beijing 100083 Peoples R China

Imbalanced class distribution of time series data often results in bias in the classification of surfaces, the classifier cannot usually achieve the best classification performance. When the classifier fails to detect a minority class of data, it indicates that the model has failed. In the real world, there are serious consequences when models fail to classify the minority classes of data (rare events). In this paper, we propose a new model that integrates a long and short-term memory (lstm)-based autoencoder (AE) network with dense weighted small spheres and large margins (lstmAE-DWSSLM) for the classification of imbalanced time series. The long short-term memory autoencoder network can learn temporally dependent feature representations from unlabeled data. The optimal feature representation can improve the classification performance. The density weight function is applied to the small spheres and large margins (SSLM) classifier to effectively handle the imbalanced classification task. Finally, two metrics are chosen to evaluate the performance of the model. Experimental results on multiple UCR datasets and EEG datasets validate the significant advantages of our proposed model. Accordingly, our model is applied to the actual datasets, good results are obtained in the plant electrical signal datasets, which further validated the practicality of the method. The comprehensive experimental results show that for imbalanced time series classification, the proposed method outperforms similar methods.

关键词： Imbalanced time series data lstm autoencoder Density weight Small Sphere and Large Margin Classification

来源：评论

学校读者我要写书评

暂无评论

The role of latent representations for design space exploration of floorplans

引用

SIMULATION-TRANSACTIONS OF THE SOCIETY FOR MODELING AND SIMULATION INTERNATIONAL 2023年第11期99卷 1167-1179页

作者： Azizi, Vahid Usman, Muhammad Sohn, Samuel S. Schwartz, Mathew Moon, Seonghyeon Faloutsos, Petros Kapadia, Mubbasir Rutgers State Univ Dept Comp Sci New Brunswick NJ USA King Fahd Univ Petr & Minerals Dept Informat & Comp Sci Dhahran 31261 Saudi Arabia New Jersey Inst Technol Coll Architecture & Design Newark NJ USA York Univ Dept Elect Engn & Comp Sci Toronto ON Canada Univ Hlth Network Toronto Rehabil Inst UHN Toronto Rehabil Inst Toronto ON Canada

Floorplans often require considering numerous factors, from the layout size to cost, numeric attributes such as room sizes, and other intrinsic properties such as connectivity between visible regions. Representing these complex factors is challenging, but doing so in a representative and efficient way can enable new modes of design exploration. Existing image and graph-based approaches of floorplans' representation often failed to consider low-level space semantics, structural features, and space utilization with respect to its future inhabitants, which are all the critical elements to analyze design layouts. We present a latent-space representation of floorplans using gated recurrent unit variational autoencoder (GRU-VAE), where floorplans are represented as attributed graphs (encoded with the abovementioned features). Two local search approaches are presented to efficiently explore the latent space for optimizing and generating new floorplans for the given environment. Semantic, structural, and visibility metrics are evaluated individually and as a combined objective for optimizations.

关键词： Floorplan representation floorplan optimization floorplan generation lstm autoencoder GRU variational autoencoder latent search space attributed graphs isovists human behavioral features

来源：评论

学校读者我要写书评

暂无评论

Predicting status of pre- and post-M&A deals using machine learning and deep learning techniques

引用

Digital Finance 2025年第1期7卷 61-106页

作者： Karatas, Tugce Hirsa, Ali Department of IEOR Columbia University NY United States

Risk arbitrage or merger arbitrage is a well-known investment strategy that speculates on the success of M&A deals. Prediction of the deal status in advance is of great importance for risk arbitrageurs. If a deal is mistakenly classified as a completed deal, then enormous cost can be incurred as a result of investing in target company shares. On the contrary, risk arbitrageurs may lose the opportunity of making profit. In this paper, we present an ML- and DL-based methodology for deal announcement prediction based on rumor and takeover success prediction problem. We initially apply various ML techniques for data preprocessing: (a) kNN for data imputation, (b) PCA for lower dimensional representation of numerical variables, (c) MCA for categorical variables, and (d) lstm autoencoder for sentiment scores. We experiment with different cost functions, different evaluation metrics, and oversampling techniques to tackle class imbalance in our dataset. We propose four novel classification frameworks that integrate the sentiment scores into the feedforward neural networks in different settings. We optimize the hyperparameter architecture of each model based on the chosen evaluation criteria with SMBO-TPE algorithm. The results show that our model frameworks outperform the benchmark models and also show robust performance over the different market environments. © The Author(s), under exclusive licence to Springer Nature Switzerland AG 2025.

关键词： KNN lstm autoencoder M& A MCA Neural networks PCA SMOTE Takeover success

来源：评论

学校读者我要写书评

暂无评论

Validation of unsupervised anomaly detection from historical satellite positional data 28

Validation of unsupervised anomaly detection from historical...

引用

28th International Conference on Information Visualisation (IV)

作者： Dracek, Frantisek Durikovic, Roman Silha, Jiri Comenius Univ Fac Math Phys & Informat Bratislava Slovakia

ISBN: (纸本)9798350380170;9798350380163

This paper presents unsupervised framework for detection of anomalies in trajectory of satellites from historical positional data in the form of Two-line element set (TLE) data. It describes the preprocessing steps tailored for orbital parameters which include smoothing with a weighted moving average, resampling to a daily frequency, and applying a Gaussian derivative filter to enhance changes. Novel angular variable preprocessing approach is employed to handle periodic discontinuities. The anomaly detection model employs a sequence-to-sequence Long Short-Term Memory (lstm) autoencoder architecture to model the time series of orbital parameters. Anomalies are identified based on the reconstruction error of the autoencoder. The framework is validated on a small dataset containing satellites with known maneuver events, demonstrating accurate detection of both large-scale maneuvers and small orbital maintenance maneuvers.

关键词： TLE lstm autoencoder Gaussian derivative angular variables validation multivariate time series anomaly detection

来源：评论

学校读者我要写书评

暂无评论

Novelty Detection of a Rolling Bearing using Long Short-Term Memory autoencoder 37

Novelty Detection of a Rolling Bearing using Long Short-Term...

引用

37th International Technical Conference on Circuits/Systems, Computers and Communications (ITC-CSCC)

作者： Asavalertpalakom, Sunithi Singhatanadgid, Pairod Ardsomang, Tutpol Chulalongkorn Univ Fac Engn Dept Mechan Engn Bangkok Thailand Digital Asset Solut Rayong Engn & Plant Serv Rayong Thailand

ISBN: (数字)9781665485593

ISBN: (纸本)9781665485593

The major challenge in predictive maintenance (PdM) is an insufficiency of failure data used to train the model and the high complexity of industrial plants, where operational conditions changed over time hence the predetermined threshold will change frequently. Novelty detection offers a solution to this problem by detecting anomalies through learning only the normal data. This study implements a novelty detection method as a one-class classification to detect early signs of failure in a rolling bearing using long short-term memory (lstm) autoencoder. lstm autoencoder is a deep learning algorithm combining an autoencoder and lstm network to reconstruct the normal data to learn nonlinear relationships and temporal nature. The model was tested on historical multivariate time-series fault data provided by Case Western Reserve University. The dropout layer is implemented to reduce overfitting by creating an ensemble model of a neural network. The results suggested that the lstm autoencoder can effectively differentiate between normal and fault patterns of the bearing with up to 92 % accuracy.

关键词： artificial intelligence predictive maintenance novelty detection deep learning lstm autoencoder rolling bearing

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：