Search results

The presented research proposes a method aimed to improve a case retrieval phase of the case-based reasoning (CBR) system through optimization of feature relevance parameters, i.e. feature weights.

The improvement is achieved by applying the metaheuristic optimization technique, called electromagnetism-like algorithm (EM), in order to appropriately adjust the feature weights used in k-NN classifier. The usability of the proposed EM k-NN algorithm is much broader since it can also be used outside the CBR system, e.g. for solving general pattern recognition tasks.

It is showed that the proposed EM k-NN algorithm improves the baseline k-NN model and outperforms the appropriately tuned artificial neural network (ANN) in the task of predicting the case (data record) output values. The results are verified by performing statistical analysis.

The proposed method is currently adjusted to deal with numerical features, so, as a direction for future work, the variant of EM k-NN algorithm that deals with symbolic or some more complex types of features should be considered.

EM k-NN algorithm can be incorporated as a case retrieval component inside a general CBR system. This is the future direction of the investigation since the authors intend to build a complete specialized CBR system for construction project management. The overall CBR with incorporated EM k-NN will have significant implication in the construction management as it will be able to produce more accurate prediction of viability and the life cycle of new construction projects.

The electromagnetism-like algorithm is applied to the problem of finding feature weights for the first time. EM potential for solving the problem of weighting features lies in its internal structure because it is based on the real-valued EM vectors. The overall EM k-NN algorithm is applied on data sets generated from real construction projects data corpus. The proposed algorithm proved its efficiency as it outperformed baseline k-NN model and ANN. Its applicability in more complex and specialized CBR systems is high since it can be easily added due to its modular (black-box) design.

Our proposed machine learning model contributes to improving the quality of Hire by providing a more nuanced and comprehensive analysis of candidate attributes. Instead of focusing solely on obvious factors, such as qualifications and experience, our model also considers various dimensions of fit, including person-job fit and person-organization fit. By integrating these dimensions of fit into the model, we can better predict a candidate’s potential contribution to the organization, hence enhancing the Quality of Hire.

Within the scope of the investigation, the competencies of the personnel working in the IT department of one in the largest state banks of the country were used. The entire data collection includes information on 1,850 individual employees as well as 13 different characteristics. For analysis, Python’s “keras” and “seaborn” modules were used. The Gower coefficient was used to determine the distance between different records.

The K-NN method resulted in the formation of five clusters, represented as a scatter plot. The axis illustrates the cohesion that exists between things (employees) that are similar to one another and the separateness that exists between things that have their own individual identities. This shows that the clustering process is effective in improving both the degree of similarity within each cluster and the degree of dissimilarity between clusters.

Employee competencies were evaluated within the scope of the investigation. Additionally, other criteria requested from the employee were not included in the application.

This study will be beneficial for academics, professionals, and researchers in their attempts to overcome the ongoing obstacles and challenges related to the securing the proper talent for an organization. In addition to creating a mechanism to use big data in the form of structured and unstructured data from multiple sources and deriving insights using ML algorithms, it contributes to the debates on the quality of hire in an entire organization. This is done in addition to developing a mechanism for using big data in the form of structured and unstructured data from multiple sources.

Hybrid systems as the next generation of intelligent applications within the field of mass appraisal and valuation are investigated. Motivated by the obvious limitations of paradigms that are being used in isolation or as stand‐alone techniques such as multiple regression analysis, artificial neural networks and expert systems. Clearly, there are distinct advantages in integrating two or more information processing systems that would address some of the discrete problems of individual techniques. Examines first, the strategic development of mass appraisal approaches which have traditionally been based on “stand‐alone” techniques; second, the potential application of an intelligent hybrid system. Highlights possible solutions by investigating various hybrid systems that may be developed incorporating a nearest neighbour algorithm (k‐NN). The enhancements are aimed at two major deficiencies in traditional distance metrics; user dependence for attribute weights and biases in the distance metric towards matching categorical variables in the retrieval of neighbours. Solutions include statistical techniques: mean, coefficient of variation and significant mean. Data mining paradigms based on a loosely coupled neural network or alternatively a tight coupling with genetic algorithms are used to discover attribute weights. The hybrid architectures developed are applied to a property data set and their performance measured based on their predictive value as well as perspicuity. Concludes by considering the application and the relevance of these techniques within the field of computer assisted mass appraisal.

Location awareness is essential to decisions pertinent to tracking and progress reporting, as well as to safety in construction projects. However, these applications have been mostly limited to the outdoor environment, where satellites for positioning information are in view. Recent studies on indoor location sensing systems are now overcoming this limitation and offering significant potential on construction practices, and radio frequency identification (RFID) is the most widely utilised technology for such application. The purpose of this paper is to address a wide range of protocols that are vital for RFID deployment for indoor construction. The paper identifies deployment settings to provide data acquisition with higher accuracy for indoor location sensing in construction.

A computational platform was designed to assess and evaluate the most suitable condition related to deployment of reference tags in construction. In this platform, a number of protocols and parameters are presented and their performance is evaluated. The evaluation scenarios were performed on a construction facility in Montreal, as well as in a controlled lab environment. The computational platform used for the study comprises the use of passive reference RFID tags and K Nearest Neighbour algorithm (K‐NN) for course‐grained detection of target's location and its classification into pre‐defined zone areas.

The studies resulted in a number of observations, findings, and lessons learned for RFID deployment in construction. The results indicate that: the speed of the reader is in direct relationship with the detection error rate; zone configuration effectiveness is in direct relationship with the deployed RFID read‐range; error rate on the controlled environment is significantly lower than rates in construction site; and stationary reader performs better than moving reader.

The paper's findings are expected to be of considerable value to researchers and practitioners involved in the utilisation of RFID technology in construction. The paper provides a set of helpful protocols for the deployment of passive RFIDs for automated onsite management of construction operations.

The purpose of optimal protocol prediction and the benefits offered by quantum key distribution (QKD), including unbreakable security, there is a growing interest in the practical realization of quantum communication. Realization of the optimal protocol predictor in quantum key distribution is a critical step toward commercialization of QKD.

The proposed work designs a machine learning model such as K-nearest neighbor algorithm, convolutional neural networks, decision tree (DT), support vector machine and random forest (RF) for optimal protocol selector for quantum key distribution network (QKDN).

Because of the effectiveness of machine learning methods in predicting effective solutions using data, these models will be the best optimal protocol selectors for achieving high efficiency for QKDN. The results show that the best machine learning method for predicting optimal protocol in QKD is the RF algorithm. It also validates the effectiveness of machine learning in optimal protocol selection.

The proposed work was done using algorithms like the local search algorithm or exhaustive traversal, however the major downside of using these algorithms is that it takes a very long time to revert back results, which is unacceptable for commercial systems. Hence, machine learning methods are proposed to see the effectiveness of prediction for achieving high efficiency.

The paper aims to develop a novel method for the classification of different physical activities of a human being, using fabric sensors. This method focuses mainly on classifying the physical activity between normal action and violent attack on a victim and verifies its validity.

The system is realized as a protective jacket that can be worn by the subject. Stretch sensors, pressure sensors and a 9 degree of freedom accelerometer are strategically woven on the jacket. The jacket has an internal bus system made of conductive fabric that connects the sensors to the Flora chip, which acts as the data acquisition unit for the data generated. Different activities such as still, standing up, walking, twist-jump-turn, dancing and violent action are performed. The jacket in this study is worn by a healthy subject. The main phases which describe the activity recognition method undertaken in this study are the placement of sensors, pre-processing of data and deploying machine learning models for classification.

The effectiveness of the method was validated in a controlled environment. Certain challenges are also faced in building the experimental setup for the collection of data from the hardware. The most tedious challenge is to collect the data without noise and error, created by voltage fluctuations when stretched. The results show that the support vector machine classifier can classify different activities and is able to differentiate normal action and violent attacks with an accuracy of 98.8%, which is superior to other methods and algorithms.

This study leads to an understanding of human physical movement under violent activity. The results show that data compared with normal physical motion, which includes even a form of dance is quite different from the data collected during violent physical motion. This jacket construction with woven sensors can capture every dimension of the physical motion adding features to the data on which the machine learning model will be built.

Unlike other studies, where sensors are placed on isolated parts of the body, in this study, the fabric sensors are woven into the fabric itself to collect the data and to achieve maximum accuracy instead of using isolated wearable sensors. This method, together with a fabric pressure and stretch sensors, can provide key data and accurate feedback information when the victim is being attacked or is in a normal state of action.

The paper aims to investigate the application of particle swarm optimisation and back propagation in weights optimisation and training of artificial neural networks within the mass appraisal industry and to compare the performance with standalone back propagation, genetic algorithm with back propagation and regression models.

The study utilised linear regression modelling before the semi-log and log-log models with a sample of 3,242 single-family dwellings. This was followed by the hybrid systems in the selection of optimal attribute weights and training of the artificial neural networks. Also, the standalone back propagation algorithm was used for the network training, and finally, the performance of each model was evaluated using accuracy test statistics.

The study found that combining particle swarm optimisation with back propagation in global and local search for attribute weights enhances the predictive accuracy of artificial neural networks. This also enhances transparency of the process, because it shows relative importance of attributes.

A robust assessment of the models’ predictive accuracy was inhibited by fewer accuracy test statistics found in the software. The research demonstrates the efficacy of combining two models in the assessment of property values.

This work demonstrated the practicability of combining particle swarm optimisation with back propagation algorithms in finding optimal weights and training of the artificial neural networks within the mass appraisal environment.

This paper conducts a Systematic Literature Review (SLR) of Machine Learning (ML) in Supply Chain Management through bibliometric and network analysis, the authors are able to grasp key features of the contemporary literature. The study makes use of state-of-the-art analytical framework based on a unified approach to reveal insights from the present body of knowledge and the potentials for future research developments.

Unlike standard literature reviews, in SLR, a structured approach is followed. The approach enables utilizing contemporary tools and software packages such as R-package “bibliometrix” and Gephi for exploratory and visual analytics. A number of clustering methods are employed to form clusters. Later, multivariate analysis methodologies are adopted to determine the dominant clusters for the influential co-cited references.

Using contemporary tools from Bibliometric Analysis (BA), the authors identify in an exploratory analysis, the influential authors, sources, regions, affiliations and papers. In addition, the use of network analysis tools reveals research clusters, topological analysis, key research topics, interrelation and authors’ collaboration along with their patterns. Finally, the optimum number of clusters computed for cluster analysis is decided using a systematic procedure based on multivariate analysis such as k-means and factor analysis.

Modern-day supply chains increasingly depend on developing superior insights from large amounts of data available from diverse sources in unstructured and semi-structured formats. In order to maintain a competitive edge, the supply chains need to perform speedy analysis of big data using efficient tools that provide real-time decision-making insights. Such an analysis necessitates automated processing using intelligent ML algorithms. Through a BA followed by a detailed data visualization in a network analysis enabled grasping key features of the contemporary literature. The analysis is based on 155 documents from the period 2008 to 2018 selected using a systematic selection procedure.

Mobile computing research has been focused on developing technologies for handheld devices such as mobile phones, notebook computers, and mobile IP. Today, emphasis is increasing on context‐aware computing, which aims to build the intelligence into mobile devices to sense and respond to the user's context. The purpose of this paper is to present a context‐aware mobile computing model (ContextAlert) that senses the user's context and intelligently configures the mobile phone alert mode accordingly.

The paper proposes a three‐step approach in designing the model based on the embedded sensor data (accelerometer, GPS antenna, and microphone) of a G1 Adriod phone. As adaptivity is essential for context‐aware computing, within this model a new learning mechanism is presented to maintain a constant adaptivity rate for new learning while keeping the catastrophic forgetting problem minimal.

The model has been evaluated in many aspects using data collected from human subjects. The experiment results show that the proposed model performs well and yields a promising result.

This paper is distinguished from other previous papers by: first, using multiple sensors embeded in the mobile phone, which is more realistic for detecting the user's context than having various sensors attached to different parts of user's body; second, by being a novel model that uses sensed contextual information to provide a service that better synchronizes the user's daily life with a context‐aware alert mode. With this service, the user can avoid the problems such as forgetting to switch to vibrate mode while in a meeting or a movie theater, and taking the risk of picking up a phone call while driving, and third, being an adaptive learning algorithm that maintains a constant adaptivity rate for new learning while keeping the catastrophic forgetting problem minimal.

This paper aims to investigate the methods for the prediction of tags on a textual corpus that describes diverse data sets based on short messages; as an example, the authors demonstrate the usage of methods based on hotel staff inputs in a ticketing system as well as the publicly available StackOverflow corpus. The aim is to improve the tagging process and find the most suitable method for suggesting tags for a new text entry.

The paper consists of two parts: exploration of existing sample data, which includes statistical analysis and visualisation of the data to provide an overview, and evaluation of tag prediction approaches. The authors have included different approaches from different research fields to cover a broad spectrum of possible solutions. As a result, the authors have tested a machine learning model for multi-label classification (using gradient boosting), a statistical approach (using frequency heuristics) and three similarity-based classification approaches (nearest centroid, k-nearest neighbours (k-NN) and naive Bayes). The experiment that compares the approaches uses recall to measure the quality of results. Finally, the authors provide a recommendation of the modelling approach that produces the best accuracy in terms of tag prediction on the sample data.

The authors have calculated the performance of each method against the test data set by measuring recall. The authors show recall for each method with different features (except for frequency heuristics, which does not provide the option to add additional features) for the dmbook pro and StackOverflow data sets. k-NN clearly provides the best recall. As k-NN turned out to provide the best results, the authors have performed further experiments with values of k from 1–10. This helped us to observe the impact of the number of neighbours used on the performance and to identify the best value for k.

The value and originality of the paper are given by extensive experiments with several methods from different domains. The authors have used probabilistic methods, such as naive Bayes, statistical methods, such as frequency heuristics, and similarity approaches, such as k-NN. Furthermore, the authors have produced results on an industrial-scale data set that has been provided by a company and used directly in their project, as well as a community-based data set with a large amount of data and dimensionality. The study results can be used to select a model based on diverse corpora for a specific use case, taking into account advantages and disadvantages when applying the model to your data.