Knowledge Discovery from Large Amounts of Social Media Data

Belcastro, Loris; Cantini, Riccardo; Marozzo, Fabrizio

doi:10.3390/app12031209

Cited by 11 publications

(9 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…After primary data cleaning and oscillation resolution [15], both datasets contain over two million mobile users and thousands of base stations. Different from the sparse social media check-in data [49], which are collected when users check in on mobile applications, the time interval between two consecutive records of one user is usually around 30 min in cellular network data. In practice, many users always stay in a street block for several hours during the day.…”

Section: Datasetmentioning

confidence: 99%

Human Mobility Prediction with Calibration for Noisy Trajectories

Miao

Lin

et al. 2022

Electronics

View full text Add to dashboard Cite

Human mobility prediction is a key task in smart cities to help improve urban management effectiveness. However, it remains challenging due to widespread intractable noises in large-scale mobility data. Based on previous research and our statistical analysis of real large-scale data, we observe that there is heterogeneity in the quality of users’ trajectories, that is, the regularity and periodicity of one user’s trajectories can be quite different from another. Inspired by this, we propose a trajectory quality calibration framework for quantifying the quality of each trajectory and promoting high-quality training instances to calibrate the final prediction process. The main module of our approach is a calibration network that evaluates the quality of each user’s trajectories by learning their similarity between them. It is designed to be model-independent and can be trained in an unsupervised manner. Finally, the mobility prediction model is trained with the instance-weighting strategy, which integrates quantified quality scores into the parameter updating process of the model. Experiments conducted on two citywide mobility datasets demonstrate the effectiveness of our approach when dealing with massive noisy trajectories in the real world.

show abstract

Section: Datasetmentioning

confidence: 99%

Human Mobility Prediction with Calibration for Noisy Trajectories

Miao

Lin

et al. 2022

Electronics

View full text Add to dashboard Cite

show abstract

“…In the application developed to discover interesting social media topics, tweets related to Covid-19 were handled and it was seen that the developed application achieved the best results compared to existing models in discovering main hashtag-based topics with 0.77 accuracy. 13 As neighborhood appreciation became a recent trend for researchers:…”

Section: Sociologymentioning

confidence: 99%

“…In the application developed to discover the polarity of public opinion, 820,000 tweets about the 2016 U.S. presidential election were used, and the winning candidate was correctly identified in 8 of 10 states. In the application developed to discover interesting social media topics, tweets related to Covid‐19 were handled and it was seen that the developed application achieved the best results compared to existing models in discovering main hashtag‐based topics with 0.77 accuracy 13 …”

Section: Research Field‐based Literature Reviewmentioning

confidence: 99%

Social big data applications and challenges

Erol¹,

Aksoy²,

Sağıroğlu³

2022

Concurrency and Computation

View full text Add to dashboard Cite

Lifestyles of individuals have changed drastically in the last two decades with the impact of social media platforms which transforms individuals from being users into an asset of social media. The assets now become very precious and seriously attract who can generate useful or harmful values. In this context, studies conducted in the last 5 years are analyzed based on the methodology covering implementation areas, data sources, data size, methods and tools. The studies were classified and summarized under nine main "research fields," and a "purpose-based" classification under three main purposes was investigated. The results have shown that even if data obtained from social media platforms are often preferred in the studies, issues such as compliance with legal regulations, data processing, confidentiality and privacy of data also bring difficulties; collection and processing of social big data are a serious obstacle to the realization of many studies; not enough data sources provided by public or private enterprises; most of the studies carried out on text data, and the rest focused on location and image data; mostly machine learning methods are preferred in applications. This study differs from previous literature reviews by revealing comprehensively how social big data can be transformed into practice with a holistic perspective.

show abstract

“…The fourth paper [15] presents three social big data analysis applications, defined and executed in parallel on a cloud platform by using ParSoDA [16], a programming library written in Java that enables developers to create cloud-based parallel applications for analyzing large volumes of social media data. Such applications focused on analyzing data from three different perspectives: (i) discovering the main tourist attractions and also the mobility patterns (i.e., trajectories) from geotagged posts [17]; (ii) understanding the political orientation of social media users so as to predict the outcome of political events [18]; (iii) analyzing the hashtags used by social media users to discover the main topics underlying social media conversation and how users refer to them in publishing online content [19].…”

Section: Cloud Computing For Big Data Analysismentioning

confidence: 99%