Deep LSTM Recurrent Neural Networks for Arterial Traffic Volume Data Imputation

2021 ◽  
Author(s):  
Abhilasha J. Saroj ◽  
Angshuman Guin ◽  
Michael Hunter
Keyword(s):  
Neural Networks ◽  
Recurrent Neural Networks ◽  
Traffic Volume ◽  
Data Imputation ◽  
Volume Data ◽  
Arterial Traffic

scholarly journals A Kriging based spatiotemporal approach for traffic volume data imputation

PLoS ONE ◽  
2018 ◽  
Vol 13 (4) ◽  
pp. e0195957 ◽  
Author(s):  
Hongtai Yang ◽  
Jianjiang Yang ◽  
Lee D. Han ◽  
Xiaohan Liu ◽  
Li Pu ◽  
...  
Keyword(s):  
Traffic Volume ◽  
Data Imputation ◽  
Volume Data

Analisis Kinerja dan Pengaruh Pelebaran Jalan serta Pembuatan Median Jalan (Studi Kasus Jalan Temanggung Tilung Kota Palangkaraya)

2018 ◽  
Vol 7 (1) ◽  
pp. 51-60
Author(s):  
Fitri Wulandari ◽  
Nirwana Puspasari ◽  
Noviyanthy Handayani
Keyword(s):  
Data Retrieval ◽  
Service Level ◽  
Traffic Volume ◽  
Primary Data ◽  
Volume Data ◽  
Saturation Degree ◽  
The Road ◽  
Before And After ◽  
Road Widening ◽  
The Way

Jalan Temanggung Tilung is a 2/2 UD type road (two undirected two-way lanes) with a road width of 5.5 meters, which is a connecting road between two major roads, namely the RTA road. Milono and the path of G. Obos. Over time, the volume of traffic through these roads increases every year, plus roadside activities that also increase cause congestion at several points of the way. To overcome this problem, the local government carried out road widening to increase the capacity and level of road services. The study was conducted to determine the amount of traffic volume, performance, service level of the Temanggung Tilung road section at peak traffic hours before and after road widening. Data retrieval is done by the direct survey to the field to obtain primary data in the form of geometric road data, two-way traffic volume data, and side obstacle data. Performance analysis refers to the 1997 Indonesian Road Capacity Manual (MKJI) for urban roads. From the results of data processing, before increasing the road (Type 2/2 UD), the traffic volume that passes through the path is 842 pcs/hour and after road widening (Type 4/2 UD) the traffic volume for two directions is 973 pcs/hour, with route A equaling 528 pcs/hour and direction B equaling 445 pcs/hour. Based on the analysis of road performance before road enhancement, the capacity = 2551 pcs/hour, saturation degree = 0.331, and the service level of the two-way road are level B. Based on the analysis of the performance of the way after increasing the way, the direction capacity A = 2686 pcs/hour and direction B = 2674 pcs /hour, saturation degree for direction A = 0.196 and direction B = 0.166, service level for road direction A and direction B increase to level A

Levenshtein Augmentation Improves Performance of SMILES Based Deep-Learning Synthesis Prediction

2020 ◽  
Author(s):  
Dean Sumner ◽  
Jiazhen He ◽  
Amol Thakkar ◽  
Ola Engkvist ◽  
Esben Jannik Bjerrum
Keyword(s):  
Neural Networks ◽  
Pattern Recognition ◽  
Deep Learning ◽  
Recurrent Neural Networks ◽  
Data Augmentation ◽  
State Of The Art ◽  
Sequence Similarity ◽  
Learning Models ◽  
Underlying Network

<p>SMILES randomization, a form of data augmentation, has previously been shown to increase the performance of deep learning models compared to non-augmented baselines. Here, we propose a novel data augmentation method we call “Levenshtein augmentation” which considers local SMILES sub-sequence similarity between reactants and their respective products when creating training pairs. The performance of Levenshtein augmentation was tested using two state of the art models - transformer and sequence-to-sequence based recurrent neural networks with attention. Levenshtein augmentation demonstrated an increase performance over non-augmented, and conventionally SMILES randomization augmented data when used for training of baseline models. Furthermore, Levenshtein augmentation seemingly results in what we define as <i>attentional gain </i>– an enhancement in the pattern recognition capabilities of the underlying network to molecular motifs.</p>

Sign in / Sign up

Export Citation Format

Share Document