Filters








132 Hits in 5.4 sec

Learning Interpretable Feature Context Effects in Discrete Choice [article]

Kiran Tomlinson, Austin R. Benson
2020 arXiv   pre-print
However, identifying these effects from observed choices is challenging, often requiring foreknowledge of the effect to be measured.  ...  Using our models, we identify new context effects in widely used choice datasets and provide the first analysis of choice set context effects in social network growth.  ...  Identifying choices from temporal network data. Our network analysis assumes that the graphs grow according to a multimode model that combines triadic closure with a method of global edge formation.  ... 
arXiv:2009.03417v2 fatcat:qpadn6n65rgfdiucgqaqwxpsli

Theory-based residual neural networks: A synergy of discrete choice models and deep neural networks [article]

Shenhao Wang, Baichuan Mo, Jinhua Zhao
2020 arXiv   pre-print
The TB-ResNet framework is simple, as it uses a (δ, 1-δ) weighting to take advantage of DCMs' simplicity and DNNs' richness, and to prevent underfitting from the DCMs and overfitting from the DNNs.  ...  This framework is also flexible: three instances of TB-ResNets are designed based on multinomial logit model (MNL-ResNets), prospect theory (PT-ResNets), and hyperbolic discounting (HD-ResNets), which  ...  DNNs generate dominant predictive performance because they can automatically learn utility specification, termed as an "end-to-end" system that can "learn from scratch".  ... 
arXiv:2010.11644v1 fatcat:6a7k443jtfgjlpqf577xajuz3e

A survey of preference estimation with unobserved choice set heterogeneity

Gregory S. Crawford, Rachel Griffith, Alessandro Iaria
2020 Journal of Econometrics  
of MNL models from subsets of true and observed choice sets.  ...  However, it enables researchers to learn about both consumer preferences and choice set formation.  ...  Importance sampling procedure from Goeree (2008) In this Appendix, we detail how to implement the importance sampling procedure proposed by Goeree (2008) for the estimation of model (3.2) when choice  ... 
doi:10.1016/j.jeconom.2020.07.024 fatcat:vskep7yplbdg7e5njhdcuzn55q

Review on Machine Learning Techniques for Developing Pavement Performance Prediction Models

Rita Justo-Silva, Adelino Ferreira, Gerardo Flintsch
2021 Sustainability  
This work aims to review the modeling techniques that are commonly used in the development of these models. The pavement deterioration process is stochastic by nature.  ...  It requires complex deterministic or probabilistic modeling techniques, which will be presented here, as well as the advantages and disadvantages of each of them.  ...  Abbreviations The following abbreviations are used in this manuscript: PPPMs Pavement performance prediction models PMSs Pavement management systems ML Machine learning SL Supervised learning UL Unsupervised  ... 
doi:10.3390/su13095248 fatcat:rdfr37loirgzdmzhmbckldlat4

A disaggregate freight transport chain choice model for Europe

Anders Fjendbo Jensen, Mikkel Thorhauge, Gerard de Jong, Jeppe Rich, Thijs Dekker, Daniel Johnson, Manuel Ojeda Cabral, John Bates, Otto Anker Nielsen
2019 Transportation Research Part E: Logistics and Transportation Review  
general cargo.  ...  This licence only allows you to download this work and share it with others as long as you credit the authors, but you can't change the article in any way or use it commercially.  ...  In application, freight models are almost always aggregated as they are linked with aggregated trademodels and aggregate network assignment models, and the final outputs of a freight transport model are  ... 
doi:10.1016/j.tre.2018.10.004 fatcat:me2qkg6fifbyzfkzdvhstk46nu

Modelling travellers' risky choice in a revealed preference context: a comparison of EUT and non-EUT approaches

Guotao Hu, Aruna Sivakumar, John W. Polak
2012 Transportation  
To date, however, there is little evidence to show whether the complexity of non-EUT actually leads to better model performance.  ...  The non-EUT approaches modelled in the thesis consist of Subjective Expected Value Theory, Subjective Expected Utility Theory, Weighted Utility theory, Rank Dependent Expected Value, Rank Dependent Expected  ...  These unappealing results may be partially due to the relatively low quality of network data used to generate the travel time distributions in this research. Table 5 . 5 3.  ... 
doi:10.1007/s11116-012-9408-7 fatcat:tbqkyyf5izdovnq4jmgoehjzja

The Vessel Schedule Recovery Problem (VSRP) – A MIP model for handling disruptions in liner shipping

Berit D. Brouer, Jakob Dirksen, David Pisinger, Christian E.M. Plum, Bo Vaaben
2013 European Journal of Operational Research  
predict the optimum time of van deliveries.Factors influencing the delivery process are depicted from various researches.Prediction is modeled as an artificial neural network using Back Propagation.The  ...  In considering the interests of users and investors, the proposed model attempts to determine the number and locations of bike stations, the network structure of bike paths connected between the stations  ...  We model the consumer choice behavior using a nested-MNL-based choice model, and derive a simple metric to rank products, which reveals the optimal assortment quite efficiently.  ... 
doi:10.1016/j.ejor.2012.08.016 fatcat:c27kagfnxnhjfbil2rydhjhomm

Tensor decompositions and algorithms, with applications to tensor learning [article]

Felipe Bottega Diniz
2021 arXiv   pre-print
In chapter 5 we consider the intersection between tensor decompositions and machine learning. A novel model is introduced, which works as a tensor version of neural networks.  ...  We begin with some examples of CPD applications to real world problems. A short summary of the main contributions in this work follows.  ...  Tensor learning: (3LmnR + 2L 2 mnR 2 + L 2 mnR)N E flops Neural network: ((n 2 + n)C(L − 1) + mnL + m)N E flops We expect to have m, R n and L <L so the cost of the tensor learning is, in general, much  ... 
arXiv:2110.05997v1 fatcat:jktyucn73rgurg52kh4itgynmi

Moderating Effect of Alliance Network on the Relationship between Competitive Strategies and Firm Performance in the Telecommunication Industry in Kenya

Musyoka Margaret Ndunge, Robert Arasa, Charles Ombuki
2019 Zenodo  
Therefore, this has resulted in formation of alliance networks with the aim of entering new markets, developing new products faster, and meeting the growing market demands.  ...  Alliance networks enable firms respond to technological changes with greater efficiency and speed in order to remain relevant to the market scene.  ...  alliance networks as shown in model 3.2 from model 3.1, the R squared increased from 77% (table 4.23), to 81% (table 4.2), this recorded an R squared change of 4%.  ... 
doi:10.5281/zenodo.3514708 fatcat:uwcx3ao6ufh7zfwzl4iwikldha

The physics of traffic and regional development

Dirk Helbing, Kai Nagel
2004 Contemporary physics (Print)  
This contribution summarizes and explains various principles from physics which are used for the simulation of traffic flows in large street networks, the modeling of destination, transport mode, and route  ...  One can also describe daily activity patterns based on decision models, simulate migration streams, and model urban growth as a particular kind of aggregation process.  ...  From a practical point of view, the next problem then becomes to generate the destinations of the travelers.  ... 
doi:10.1080/00107510410001715944 fatcat:yyuuih3575av3anu5j6fwgzsde

Single-Pass PCA of Large High-Dimensional Data [article]

Wenjian Yu, Yu Gu, Jian Li, Shenghua Liu, Yaohang Li
2017 arXiv   pre-print
., the singular vectors corresponding to a number of dominant singular values of the data matrix) becomes a challenging task.  ...  For a set of high-dimensional data stored as a 150 GB file, the proposed algorithm is able to compute the first 50 principal components in just 24 minutes on a typical 24-core computer, with less than  ...  Introduction Many existing machine learning models, no matter supervised or unsupervised, rely on dimension reduction of input data.  ... 
arXiv:1704.07669v1 fatcat:s664uj5g25hyjhhuvnjmc6xc4i

STUD 2019 Blank Page

2019 2019 First International Conference on Smart Technology & Urban Development (STUD)  
on the readiness of information technology was in good lever, average at 4�06� Regarding the access to internet network of the university, the respondents agreed that it was in good level (4�11) with  ...  Nakhon to support educational communication in AEC [1]� This quantitative research used questionnaire to collect data from the sample group, who was 40 professors from Faculty of Business Administration  ...  With the transfer learning technique, the model is able to learn the low-level features from the whole body. Partial information around local regions is essential to distinguish persons.  ... 
doi:10.1109/stud49732.2019.9018735 fatcat:w2isrk4dsfg3fneffmbkwv5c7m

amj – Progress and Prospects

Paul Patterson, Mark D. Uncles
2005 Australasian Marketing Journal  
We present empirical applications of RBL, and we discuss its relationships to several classical models.  ...  Further, we find that there is no form of loyalty that consistently predicts all the different loyalty outcomes and, therefore, we should abandon the idea of a general concept of loyalty.  ...  The content of the book serves to introduce students to the latest research and developments in the services sector, ranging from customer relationship management, customer asset management, and six sigma  ... 
doi:10.1016/s1441-3582(05)70072-0 fatcat:cpf6yteu4nhnxbkjrxpjmazzoe

Single-Pass PCA of Large High-Dimensional Data

Wenjian Yu, Yu Gu, Jian Li, Shenghua Liu, Yaohang Li
2017 Proceedings of the Twenty-Sixth International Joint Conference on Artificial Intelligence  
., the top singular vectors of the data matrix) becomes a challenging task. In this work, a single-pass randomized algorithm is proposed to compute PCA with only one pass over the data.  ...  For a set of high-dimensional data stored as a 150 GB file, the algorithm is able to compute the first 50 principal components in just 24 minutes on a typical 24-core computer, with less than 1 GB memory  ...  Portions of the research in this paper use the FERET database of facial images collected under the FERET program, sponsored by the DOD Counterdrug Technology Development Program Office.  ... 
doi:10.24963/ijcai.2017/468 dblp:conf/ijcai/YuGLLL17 fatcat:yuzdw2tt4jbmtac7xqialyvpf4

Multiagent task allocation in social networks

Mathijs M. de Weerdt, Yingqian Zhang, Tomas Klos
2011 Autonomous Agents and Multi-Agent Systems  
Three different types of networks, namely small-world, random and scalefree networks, are used to represent various social relationships among agents in realistic applications.  ...  The results demonstrate that our algorithm works well and also that it scales well to large-scale applications.  ...  Acknowledgments This work is supported by the Technology Foundation STW, applied science division of NWO, and the Ministry of Economic Affairs.  ... 
doi:10.1007/s10458-011-9168-3 fatcat:vaajjquolrgl7pa6eh4wpdoyiy
« Previous Showing results 1 — 15 out of 132 results