BikeshareGAN: Predicting dockless bike-sharing demand based on satellite image

Journal of Transport Geography, Год журнала: 2025, Номер 126, С. 104245 - 104245

Опубликована: Апрель 28, 2025

Язык: Английский

---

A deep multi-scale neural networks for crime hotspot mapping prediction

Computers Environment and Urban Systems, Год журнала: 2024, Номер 109, С. 102089 - 102089

Опубликована: Фев. 17, 2024

Язык: Английский

---

An Optimized Machine Learning and Big Data Approach to Crime Detection

Wireless Communications and Mobile Computing, Год журнала: 2021, Номер 2021(1)

Опубликована: Янв. 1, 2021

Crime detection is one of the most important research applications in machine learning. Identifying and reducing crime rates crucial to developing a healthy society. Big Data techniques are applied collect analyse data: determine required features prime attributes that cause emergence hotspots. The traditional learning‐based algorithms lack ability generate key from dataset, hence often fail predict patterns successfully. This paper aimed at extracting such as time zones, probability, hotspots performing vulnerability analysis increase accuracy subject learning algorithm. We implemented our proposed methodology using two standard datasets. Results show feature generation method increased performance models. highest 97.5% was obtained when Naïve Bayes algorithm while analysing San Francisco dataset.

Язык: Английский

---

Criminal incidences in relation to built environment in Arba Minch City, Southern Ethiopian

SN Social Sciences, Год журнала: 2024, Номер 4(5)

Опубликована: Май 13, 2024

Язык: Английский

---

Adaptive interior design method for different MBTI personality types based on generative artificial intelligence

Architectural Intelligence, Год журнала: 2024, Номер 3(1)

Опубликована: Июль 1, 2024

Abstract Accurately predicting homeowners’ aesthetic preferences is crucial in interior design. This study develops a fine-tuning model (LORA) for design styles corresponding to different MBTI personality types, leveraging the Stable Diffusion Web UI platform and integrating it into generative artificial intelligence framework. Subsequently, personalized preference architectural renderings are recommended based on traits, aiming achieve an adaptive approach. To more precise solutions, this research surveys style color tendencies of respondents with types adds description prompts assist image generation. The finds that method can better predict favored by certain types. contributes addressing biases between designers homeowners, bringing innovative ideas methods design, expected enhance satisfaction.

Язык: Английский

---

Forecasting land surface drought in urban environments based on machine learning model

Sustainable Cities and Society, Год журнала: 2024, Номер unknown, С. 106048 - 106048

Опубликована: Дек. 1, 2024

Язык: Английский

---

Crime Prediction Methods Based on Machine Learning: A Survey

Computers, materials & continua/Computers, materials & continua (Print), Год журнала: 2022, Номер 74(2), С. 4601 - 4629

Опубликована: Окт. 31, 2022

The objective of crime prediction, one the most important technologies in social computing, is to extract useful information from many existing criminal records predict next process-related crime. It can aid police obtaining and warn public be vigilant certain areas. With rapid growth big data, Internet Things, other technologies, as well increasing use artificial intelligence forecasting models, prediction models based on deep learning techniques are accelerating. Therefore, it necessary classify algorithms compare depth attributes conditions that play an essential role analysis algorithms. Existing methods roughly divided into two categories: those conventional machine contemporary learning. This survey analyses fundamental theories procedures. frequently used data sets then enumerated, procedures various also analyzed this paper. In light insufficient scale field, ambiguity types crimes, absence have a significant impact research algorithm proposes construction learning-based model address these issues. Future researchers who will enter field provided with guide direction future development.

Язык: Английский

---

Hybrid Intelligence

Computational design and robotic fabrication, Год журнала: 2023, Номер unknown

Опубликована: Янв. 1, 2023

This open access book collects selected papers from the 4th International Conference on Computational Design and Robotic Fabrication (CDRF 2022).

Язык: Английский

---

Convolutional neural network and unmanned aerial vehicle‐based public safety framework for human life protection

International Journal of Communication Systems, Год журнала: 2023, Номер unknown

Опубликована: Май 23, 2023

Summary In this paper, we developed an object detection and identification framework to bolster public safety. Before developing the proposed framework, several existing frameworks were analyzed The other models carefully observed for their strengths weaknesses based on machine learning deep algorithms they operate on. All these kept in mind during development of model. consists unmanned aerial vehicle (UAV) utilized data collection that constantly monitors captures images designated areas. A convolutional neural network (CNN) model is recognize a threat identifies various handheld objects, such as guns knives, which facilitate criminals commit crimes. CNN comprises 16 layers with input, convolutional, dense, max‐pool, flattened different dimensions. For that, benchmarked dataset, is, small objects handled similarly weapon (SOHAs), dataset used. It six classes 8945 images, 5947 used training, 1699 testing, 849 validation. Once accomplishes classification, person criminal or non‐criminal, forwarded law enforcement agencies non‐criminal are again improvising its accuracy rate. As result, outperforms pre‐trained 0.8352 validation 0.7758. addition, gives minimal loss 0.83 0.97. decreases burden crime‐fighting increases crime detection. Additionally, it ensures fairness operates at meager computational cost compared similar models.

Язык: Английский

---

Artificial Intelligence Prediction of Urban Spatial Risk Factors from an Epidemic Perspective

Опубликована: Янв. 1, 2023

Abstract From the epidemiological perspective, previous research methods of COVID-19 are generally based on classical statistical analysis. As a result, spatial information is often not used effectively. This paper uses image-based neural networks to explore relationship between urban risk and distribution infected populations, design facilities. We take Spatio-temporal data people with new coronary pneumonia before February 28 in Wuhan 2020 as object. use kriging interpolation technology core density estimation establish epidemic heat fine grid units. further examine nine main factors, including agencies, hospitals, park squares, sports fields, banks, hotels, Etc., which tested for significant positive correlation epidemic. The weights factors training Generative Adversarial Network models, predict outbreak given area. According trained model, optimizing relevant environment areas control effectively prevents manages from dispersing. input image machine learning model city plan converted by public infrastructures, output map

Язык: Английский

---