Search results
1 – 10 of 79Ahmet Yucel, Musa Caglar, Hamidreza Ahady Dolatsara, Benjamin George and Ali Dag
Machine learning algorithms are useful to effectively analyse, and therefore automatically classify online reviews. The purpose of this paper is to demonstrate a novel text-mining…
Abstract
Purpose
Machine learning algorithms are useful to effectively analyse, and therefore automatically classify online reviews. The purpose of this paper is to demonstrate a novel text-mining framework and its potential for use in the classification of unstructured hotel reviews.
Design/methodology/approach
Well-known data mining methods (i.e. boosted decision trees (BDT), classification and regression trees (C&RT) and random forests (RF)) in conjunction with incorporating five-fold cross-validation are used to predict the star rating of the hotel reviews. To achieve this goal, extracted features are used to create a composite variable (CV) to deploy into machine learning algorithms as the main feature (variable) during the learning process.
Findings
BDT outperformed the other alternatives in the exact accuracy rate (EAR) and multi-class accuracy rate (MCAR) by reaching the accuracy rates of 0.66 and 0.899, respectively. Moreover, phrases such as “clean”, “friendly”, “nice”, “perfect” and “love” are shown to be associated with four and five stars, whereas, phrases such as “horrible”, “never”, “terrible” and “worst” are shown to be associated with one and two-star hotels, as it would be the intuitive expectation.
Originality/value
To the best of the knowledge, there is no study in the existent literature, which synthesizes the knowledge obtained from individual features and uses them to create a single composite variable that is powerful enough to predict the star rates of the user-generated reviews. This study believes that the proposed method also provides policymakers with a unique window in the thoughts and opinions of individual users, which may be used to augment the current decision-making process.
Details
Keywords
Murtaza Nasir, Carole South-Winter, Srini Ragothaman and Ali Dag
The purpose of this paper is to formulate a framework to construct a patient-specific risk score and therefore to classify these patients into various risk groups that can be used…
Abstract
Purpose
The purpose of this paper is to formulate a framework to construct a patient-specific risk score and therefore to classify these patients into various risk groups that can be used as a decision support mechanism by the medical decision makers to augment their decision-making process, allowing them to optimally use the limited resources available.
Design/methodology/approach
A conventional statistical model (logistic regression) and two machine learning-based (i.e. artificial neural networks (ANNs) and support vector machines) data mining models were employed by also using five-fold cross-validation in the classification phase. In order to overcome the data imbalance problem, random undersampling technique was utilized. After constructing the patient-specific risk score, k-means clustering algorithm was employed to group these patients into risk groups.
Findings
Results showed that the ANN model achieved the best results with an area under the curve score of 0.867, while the sensitivity and specificity were 0.715 and 0.892, respectively. Also, the construction of patient-specific risk scores offer useful insights to the medical experts, by helping them find a trade-off between risks, costs and resources.
Originality/value
The study contributes to the existing body of knowledge by constructing a framework that can be utilized to determine the risk level of the targeted patient, by employing data mining-based predictive approach.
Details
Keywords
Burak Cankaya, Berna Eren Tokgoz, Ali Dag and K.C. Santosh
This paper aims to propose a machine learning-based automatic labeling methodology for chemical tanker activities that can be applied to any port with any number of active tankers…
Abstract
Purpose
This paper aims to propose a machine learning-based automatic labeling methodology for chemical tanker activities that can be applied to any port with any number of active tankers and the identification of important predictors. The methodology can be applied to any type of activity tracking that is based on automatically generated geospatial data.
Design/methodology/approach
The proposed methodology uses three machine learning algorithms (artificial neural networks, support vector machines (SVMs) and random forest) along with information fusion (IF)-based sensitivity analysis to classify chemical tanker activities. The data set is split into training and test data based on vessels, with two vessels in the training data and one in the test data set. Important predictors were identified using a receiver operating characteristic comparative approach, and overall variable importance was calculated using IF from the top models.
Findings
Results show that an SVM model has the best balance between sensitivity and specificity, at 93.5% and 91.4%, respectively. Speed, acceleration and change in the course on the ground for the vessels are identified as the most important predictors for classifying vessel activity.
Research limitations/implications
The study evaluates the vessel movements waiting between different terminals in the same port, but not their movements between different ports for their tank-cleaning activities.
Practical implications
The findings in this study can be used by port authorities, shipping companies, vessel operators and other stakeholders for decision support, performance tracking, as well as for automated alerts.
Originality/value
This analysis makes original contributions to the existing literature by defining and demonstrating a methodology that can automatically label vehicle activity based on location data and identify certain characteristics of the activity by finding important location-based predictors that effectively classify the activity status.
Details
Keywords
Ram Jiwari and Ali Saleh Alshomrani
The main aim of the paper is to develop a new B-splines collocation algorithm based on modified cubic trigonometric B-spline functions to find approximate solutions of nonlinear…
Abstract
Purpose
The main aim of the paper is to develop a new B-splines collocation algorithm based on modified cubic trigonometric B-spline functions to find approximate solutions of nonlinear parabolic Burgers’-type equations with Dirichlet boundary conditions.
Design/methodology/approach
A modification is made in cubic trigonometric B-spline functions to handle the Dirichlet boundary conditions and an algorithm is developed with the help of modified cubic trigonometric B-spline functions. The proposed algorithm reduced the Burgers’ equations into a system of first-order nonlinear ordinary differential equations in time variable. Then, strong stability preserving Runge-Kutta 3rd order (SSP-RK3) scheme is used to solve the obtained system.
Findings
A different technique based on modified cubic trigonometric B-spline functions is proposed which is quite different from to the schemes developed in Abbas et al. (2014) and Nazir et al. (2016), and the developed algorithms are free from linearization process and finite difference operators.
Originality/value
To the best knowledge of the authors, this technique is novel for solving nonlinear partial differential equations, and the new proposed technique gives better results than the results discussed in Ozis et al. (2003), Kutluay et al. (1999), Khater et al. (2008), Korkmaz and Dag (2011), Kutluay et al. (2004), Rashidi et al. (2009), Mittal and Jain (2012), Mittal and Jiwari (2012), Mittal and Tripathi (2014), Xie et al. (2008) and Kadalbajoo et al. (2005).
Details
Keywords
The purpose of this paper is to illustrate how the numerical solution of the Burgers' equation is obtained using the methods of cubic B‐spline collocation and quadratic B‐spline…
Abstract
Purpose
The purpose of this paper is to illustrate how the numerical solution of the Burgers' equation is obtained using the methods of cubic B‐spline collocation and quadratic B‐spline Galerkin over the geometrically graded mesh.
Design/methodology/approach
The spatial domain is partitioned into geometrically graded mesh. The finite element methods are constructed within the Galerkin and collocation methods using an expansion of the quadratic and cubic B‐splines as an approximate function, respectively, over this mesh.
Findings
When the higher errors are observed at near boundaries for shock‐like and travelling wave solutions of the Burgers' equation, accuracy of the defined methods increase by using finer mesh at near this boundary.
Originality/value
Over the geometrically graded mesh definitions of the quadratic B‐spline Galerkin and cubic B‐spline collocation are given.
Details
Keywords
The purpose of this paper is to investigate the numerical solutions of the Burgers' and modified Burgers' equation using sextic B‐spline collocation method.
Abstract
Purpose
The purpose of this paper is to investigate the numerical solutions of the Burgers' and modified Burgers' equation using sextic B‐spline collocation method.
Design/methodology/approach
Crank‐Nicolson central differencing scheme has been used for the time integration and sextic B‐spline functions have been used for the space integration to the modified and time splitted modified Burgers' equation.
Findings
It has been found that the proposed method is unconditionally stable and obtained results are consistent with some earlier published studies.
Originality/value
Sextic B‐spline collocation method for the Burgers' and modified Burgers' equation is given.
Details
Keywords
İdris Dağ, Aynur Canivar and Ali Şahin
The purpose of this paper is to provide numerical solutions of the time‐dependent advection‐diffusion problem by using B‐spline finite element methods in which Taylor series…
Abstract
Purpose
The purpose of this paper is to provide numerical solutions of the time‐dependent advection‐diffusion problem by using B‐spline finite element methods in which Taylor series expansion is used for the related time discretization.
Design/methodology/approach
The solution domain is partitioned into uniform mesh. The collocation and the Galerkin methods where B‐spline functions are used as base functions are applied to advection‐diffusion equation.
Findings
Given methods are unconditionally stable and the obtained results are comparable with some earlier studies in terms of accuracy.
Originality/value
Quadratic and cubic B‐spline base functions are used with Taylor series expansion for the discretization of the equation.
Details
Keywords
Okechukwu Nwadigo, Nicola Naismith Naismith, Ali Ghaffarianhoseini, Amirhosein Ghaffarian Hoseini and John Tookey
A construction project is complex and requires dynamic modelling of a range of factors that deters time performance because of uncertainty and varying operating conditions. In…
Abstract
Purpose
A construction project is complex and requires dynamic modelling of a range of factors that deters time performance because of uncertainty and varying operating conditions. In construction project systems, the system components are the interconnected stages, which are time-dependent. Within the project stages are the activities which are the subsystems of the system components, causing a challenge to the analysis of the complex system. The relationship of construction project time management (CTM) with the construction project time influencing factors (CTFs) and the adaptability of the time-varying system is a key part of project effectiveness. This study explores the relationship between CTM and CTF, including the potentials to add dynamical changes on every project stage.
Design/methodology/approach
This study proposed a dynamic Bayesian network (DBN) model to examine the relationship between CTM and CTF. The model investigates the time performance of a construction project that enhances decision-making. First, the paper establishes a model of probabilistic reasoning and directed acrylic graph (DAG). Second, the study tests the dynamic impact (IM) of CTM-CTF on the project stages over a specific time, including the adaptability of time performance during disruptive CTF events. In demonstrating the effectiveness of the model, the authors selected one-organisation-single-location road-improvement project as the case study. Next, the confirmation of the model internal validity relied on conditional probabilities and the project knowledge experts' selected from the case company.
Findings
The study produced structural dependencies of CTM and CTF with probability observations at each stage. A predictive time performance analysis of the model at different scenarios evaluates the adaptability of CTM during CTF uncertain events. The case demonstration of the model application shows that CTFs have effects on CTM strategy, creating the observations to help time performance restorations after disruptions.
Research limitations/implications
Although the case company experts' panel confirms the internal validity of the results for managing time, the model used conditional probability table (CPT) and project state values from a project contract. A project-wide application then will require multi-case data and data-mining process for generating the CPTs.
Practical implications
The study developed a method for evaluating both quantitative and qualitative relationships between CTM and CTF, besides the knowledge to enhance CTM practice and research. In construction, the project team can use model observations to implement time performance restorations after a predictive or reactive disruption, which enhances decision-making.
Originality/value
The model used qualitative and qualitative data of a complex system to generate results, bounded by a range of probability distributions for CTM-CTF interconnections during time performance disruptions and restorations. The research explores the approach that can complement the mental CTM-CTF modeling of the project team. The CTM-CTF relationship model developed in this research is fundamental knowledge for future research, besides the valuable insight into CTF influence on CTM.
Details
Keywords
Syamsul Anwar, Taufik Djatna, Sukardi and Prayoga Suryadarma
Supply chain risks (SCRs) have uncertainty and interdependency characteristics that must be incorporated into the risk assessment stage of the SCR management framework. This study…
Abstract
Purpose
Supply chain risks (SCRs) have uncertainty and interdependency characteristics that must be incorporated into the risk assessment stage of the SCR management framework. This study aims to develop SCR networks and determine the major risk drivers that impact the performance of the sago starch agro-industry (SSA).
Design/methodology/approach
The risk and performance variables were collected from the relevant literature and expert consultations. The Bayesian network (BN) approach was used to model the uncertain and interdependent SCRs. A hybrid method was used to develop the BN structure through the expert’s knowledge acquisitions and the learning algorithm application. Sensitivity analyses were performed to examine the significant risk driver and their related paths.
Findings
The analyses of model indicated several significant risk drivers that could affect the performance of the SSA. These SCR including both operational and disruption risks across sourcing, processing and delivery stage.
Research limitations/implications
The implementation of the methodology was only applied to the Indonesian small-medium size sago starch agro-industry. The generalization of findings is limited to industry characteristics. The modelled system is restricted to inbound, processing and outbound logistics with the risk perspective from the industry point of view.
Practical implications
The results of this study assist the related actors of the sago starch agro-industry in recognizing the major risk drivers and their related paths in impacting the performance measures.
Originality/value
This study proposes the use of a hybrid method in developing SCR networks. This study found the significant risk drivers that impact the performance of the sago starch agro-industry.
Details
Keywords
Sapna Pandit, Manoj Kumar, R.N. Mohapatra and Ali Saleh Alshomrani
This paper aims to find the numerical solution of planar and non-planar Burgers’ equation and analysis of the shock behave.
Abstract
Purpose
This paper aims to find the numerical solution of planar and non-planar Burgers’ equation and analysis of the shock behave.
Design/methodology/approach
First, the authors discritize the time-dependent term using Crank–Nicholson finite difference approximation and use quasilinearization to linearize the nonlinear term then apply Scale-2 Haar wavelets for space integration. After applying this scheme on partial differential, the equation transforms into a system of algebraic equation. Then, the system of equation is solved using Gauss elimination method.
Findings
Present method is the extension of the method (Jiwari, 2012). The numerical solutions using Scale-2 Haar wavelets prove that the proposed method is reliable for planar and non-planar nonlinear Burgers’ equation and yields results better than other methods and compatible with the exact solutions.
Originality/value
The numerical results for non-planar Burgers’ equation are very sparse. In the present paper, the authors identify where the shock wave and discontinuity occur in planar and non-planar Burgers’' equation.
Details