China 
Africa 
Explore chapters and articles related to this topic
Inhalable Particulate Matter and Extractable Organic Matter Receptor Source Apportionment Models for the ATEOS Urban Sites
Published in Paul J. Lioy, Joan M. Daisey, Toxic Air Pollution, 1987
Maria T. Morandi, Joan M. Daisey, Paul J. Lioy
Multiple regression apportionment models were developed using forward stepwise regression with listwise deletion of missing data and the SPSS default criteria of F = 1.00, and T = 0.001.
Out-of-Sequence Measurements
Published in Jitendra R. Raol, Ajith K. Gopal, Mobile Intelligent Autonomous Systems, 2016
Much of the work on tracking and filtering is built on the assumption that the measurements are immediately available to an agent. However, it is not difficult to conceive situations in which the measurements are subject to non-negligible delays, such as the lag between measurement and receipt is of sufficient magnitude to impact on estimation. In such situations, the classical assumption that observations are available immediately is easily violated [19]. One direct solution to the OOSM problem is simply to ignore and discard the OOSM in the tracking process more like the listwise deletion is a standard default approach for dealing with missing data in most statistical packages. This solution leads obviously to a loss of the information contained in the discarded OOSM. To avoid this drawback, several alternative methods have been proposed in the literature to deal with the OOSM problem, especially for random delays. It is also striking that most of the methods proposed to handle delayed measurements have in common that delayed measurements are always ultimately incorporated into the filtering process. In the case of time delay, one common approach is related to solving a partial differential equation and boundary condition equations which do not have an explicit solution in general [23–27]. In the case of discrete time systems (and especially for random delays), the problem has been investigated via a standard Kalman filtering [28] and by augmenting the system accordingly [28–30]. Matveev and Savkin [31] consider an iterative form of state augmentation for random delays with a random lag. Larsen et al. [32] address the OOSM problem by recalculating the filter through the delayed period. In the same context, Larsen et al. [32] further propose a measurement extrapolation approximation using past and present estimates of the KF and calculating an optimal gain for this extrapolated measurement. Thomopoulos and Zhang [17] examine the case of random delay under the name of the fixed sampling and random delay filter that is shown to be equivalent to constraining the lag to a value of 1. Alexander [33] and later Larsen et al. [32] suggest using the delayed measurements to calculate a correction term and adding this to the filter estimate. Zhang et al. [34] proposed algorithms that try to minimise the information storage in an OOSM situation. Challa et al. [9] formulated the OOSM problem in a Bayesian framework. The above methods are described in more detail in Section 11.3.
Handling incomplete and missing data in water network database using imputation methods
Published in Sustainable and Resilient Infrastructure, 2020
Golam Kabir, Solomon Tesfamariam, Jordi Hemsing, Rehan Sadiq
Most of the small to medium-sized water utilities do have some basic information on existing water mains (e.g., pipe material, date of installation or pipe age, pipe diameter, etc.), but very few of them have been maintaining thorough records of pipe breaks for longer periods (Francisque et al., 2017; Kleiner & Rajani, 1999; Wood & Lence, 2009). Because of data scarcity and incomplete or missing values in particular for small to medium-sized water utilities, it is often challenging to develop statistical water main failure prediction models (Haider et al., 2014). Moreover, most of the regression based water main failure models assumed that input is provided as a complete data matrix. In that case, the dataset having missing values (one or more), the software would delete listwise or pairwise or substitute missing values with mean values (Schafer & Olsen, 1998). Any records or data matrix having a missing value in any of the variables would be deleted in listwise deletion method. In contrast, a variable containing missing values would be deleted and the remaining variables will be used for analysis in pairwise deletion (Davey & Savla, 2010; Graham, 2012).
Investigating photovoltaic solar power output forecasting using machine learning algorithms
Published in Engineering Applications of Computational Fluid Mechanics, 2022
Yusuf Essam, Ali Najah Ahmed, Rohaini Ramli, Kwok-Wing Chau, Muhammad Shazril Idris Ibrahim, Mohsen Sherif, Ahmed Sefelnasr, Ahmed El-Shafie
To deal with missing values within the raw data set in the present study, listwise deletion is employed. This is because the missing data is regarded as missing completely at random (MCAR) according to Kang (2013), due to equipment problems or failure to meet quality assessments thresholds (Marion et al., 2014). Listwise deletion involves the complete removal of the rows of data containing missing parameters. The listwise deletion is the most optimal strategy to be used in this case study as the dataset is large and the assumption of MCAR is satisfied (Kang, 2013).
How interpersonal ties affect interorganizational trust in construction projects: role differences and cross-level effects
Published in Construction Management and Economics, 2021
Wenqian Guo, Wenxue Lu, Xinran Gao, Feifei Cai
The survey was carried out between January 2019 and June 2019 through the online and on-site distribution of the questionnaire. First, questionnaires were sent to trainees involved in project management training courses in universities and corporations. These participants were veteran practitioners who engaged in project management for a wide range of large contracting enterprises or subsidiaries of group companies in China. With the permission of these respondents, the authors distributed and collected the questionnaires on-site to obtain a better response rate and more credible responses. Second, an electronic questionnaire with the same content as the paper version was produced on the website Questionnaire Star and sent to a wider source of respondents, including project management practitioners previously interviewed by the authors and graduates who had been working in project management for years in different companies. The respondents are practitioners who are engaged in the construction projects of owners, contractors, and subcontractors, including project managers, department managers, technical engineers, contract engineers, legal staff, technical staff, and general staff. Finally, 173 online questionnaires and 190 paper questionnaires were distributed and 319 questionnaires were returned in total. As for managing missing data, listwise deletion, pairwise deletion, data imputation, and full information maximum likelihood methods are commonly adopted (Marsh 1998). The authors adopted listwise deletion, a supposedly purer approach, to manage missing data, and removed these records. Then, after removing the invalid questionnaires in which the answers to several consecutive items were the same, the data were unmatched, or the answer time in electronic questionnaires was too short (within two minutes) to be considered reasonable, 271 valid questionnaires were finally obtained among which 95 are from project leaders and 176 are from team members. The effective recovery rate was 74.7%.