Efficient Training with Heterogeneous Data Distribution

2022 ◽  
pp. 98-111
2019 ◽  
Vol 79 (9-10) ◽  
pp. 6689-6708
Author(s):  
Dong Kyun Shin ◽  
Minhaz Uddin Ahmed ◽  
Yeong Hyeon Kim ◽  
Phill Kyu Rhee

Sensors ◽  
2020 ◽  
Vol 20 (5) ◽  
pp. 1443
Author(s):  
Marius Laska ◽  
Jörg Blankenbach ◽  
Ralf Klamma

The accuracy of fingerprinting-based indoor localization correlates with the quality and up-to-dateness of collected training data. Perpetual crowdsourced data collection reduces manual labeling effort and provides a fresh data base. However, the decentralized collection comes with the cost of heterogeneous data that causes performance degradation. In settings with imperfect data, area localization can provide higher positioning guarantees than exact position estimation. Existing area localization solutions employ a static segmentation into areas that is independent of the available training data. This approach is not applicable for crowdsoucred data collection, which features an unbalanced spatial training data distribution that evolves over time. A segmentation is required that utilizes the existing training data distribution and adapts once new data is accumulated. We propose an algorithm for data-aware floor plan segmentation and a selection metric that balances expressiveness (information gain) and performance (correctly classified examples) of area classifiers. We utilize supervised machine learning, in particular, deep learning, to train the area classifiers. We demonstrate how to regularly provide an area localization model that adapts its prediction space to the accumulating training data. The resulting models are shown to provide higher reliability compared to models that pinpoint the exact position.


Information sharing among the associations is a general development in a couple of zones like business headway and exhibiting. As bit of the touchy principles that ought to be kept private may be uncovered and such disclosure of delicate examples may impacts the advantages of the association that have the data. Subsequently the standards which are delicate must be secured before sharing the data. In this paper to give secure information sharing delicate guidelines are bothered first which was found by incessant example tree. Here touchy arrangement of principles are bothered by substitution. This kind of substitution diminishes the hazard and increment the utility of the dataset when contrasted with different techniques. Examination is done on certifiable dataset. Results shows that proposed work is better as appear differently in relation to various past strategies on the introduce of evaluation parameters.


2020 ◽  
Vol 4 ◽  
pp. 97-100
Author(s):  
A.P. Pronichev ◽  

The article discusses the architecture of a system for collecting and analyzing heterogeneous data from social networks. This architecture is a distributed system of subsystem modules, each of which is responsible for a separate task. The system also allows you to use external systems for data analysis, providing the necessary interface abstraction for connection. This allows for more flexible customization of the data analysis process and reduces development, implementation and support costs.


2010 ◽  
Vol 12 (2) ◽  
pp. 194-199 ◽  
Author(s):  
Shaochun DONG ◽  
Hongwei YIN ◽  
Gang XU

2010 ◽  
Vol 11 (3) ◽  
pp. 292-298
Author(s):  
Hongjun SU ◽  
Yehua SHENG ◽  
Yongning WEN ◽  
Min CHEN

Sign in / Sign up

Export Citation Format

Share Document