•935,386 Google Maps reviews of 5010 restaurants in the UK were analysed.•This study used a combination of review scraping, sentiment, text, and logistic regression analyses.•Food, service, and ...atmosphere positively affect the odds of getting 5-star rating.•The lexical analyses of food items in 8 types of restaurant were conducted.•Alcohol and a variety of dietary options positively affect the overall customer experience.
The purpose of this research is to examine the effects of restaurant attributes and the underlying factors impacting overall customer experience within a range of different restaurant types. To understand their experiences, this study analyses online reviews of restaurants which have become important sources of customer experience data. This current research utilises a combination of quantitative analyses to examine 935,386 Google Maps reviews of 5010 restaurants in London, Birmingham, and Manchester. The authors used the VADER sentiment analysis algorithm to measure the sentiment of four key restaurant attributes: food, service, atmosphere, and value. Logistic regression was conducted to test the relationships between these attributes and a 5-star rating. Furthermore, logistic regression was used to compare the changes of odds at different star rating levels. To understand the factors that drive positive and negative reviews, the top 30 food items of 8 types of restaurants were analysed.
The purpose of the current research is to develop a methodology that can analyse online reviews using machine learning techniques in such a way that practitioners in the fields of tourism and ...destination management can understand and apply the technique to improve their attractions. This research studies the TripAdvisor reviews of tourist attractions, including beaches, islands, temples, a pedestrian street, and markets in Phuket, Thailand. In total, 65,079 online reviews were analysed using two machine learning techniques: latent Dirichlet allocation (LDA) and naïve Bayes modelling.
LDA modelling helps the researchers determine the dimensions of each type of attraction. Four dimensions were specified for beaches and islands, three dimensions for a pedestrian street and temples, and two dimensions for markets. This research also developed two practical tools – dimensional salience-valence analysis (DSVA) and lexical salience-valence analysis (LSVA) – and used them to suggest actions for the Tourism Authority of Thailand (TAT).
•65,079 TripAdvisor reviews of attractions were analysed using machine learning.•10 beaches, 12 island locations, 2 temples, a street, and 12 markets were analysed.•Latent Dirichlet allocation modelling was used to extract dimensions of attractions.•Naïve Bayes modelling was used to analyse underlying terms.
The purpose of this research is to introduce a method that utilises a combination of Google Cloud Vision AI's label detection and a topic‐modelling algorithm, latent Dirichlet allocation, to identify ...common destination images and to compare destinations worldwide. The study analyses 283,912 photos of 193 countries from Flickr.com, and 16 cognitive image attributes (CIAs) are identified. Subsequent hotspot analyses indicate the exact locations of these CIAs in three sample countries: France, the US, and Thailand. Destination marketing organisations (DMOs) can use this method to more effectively analyse and promote destinations during and after the COVID‐19 pandemic.
The purpose of the current research is to introduce a new method involving machine learning that can identify and analyse the city image dimensions (CIDs) of cities worldwide. Unlike traditional ...methods, this new method can rapidly identify city image dimensions from large sets of user-generated photos in an efficient and scalable manner, which could help city managers more effectively plan city branding strategies and city development policies. Label detection with Google Cloud Vision and dimension identification (or topic extraction) with latent Dirichlet allocation (LDA) modelling were used to analyse 222,000 photos of 222 cities worldwide from Flickr.com. Theoretically, this study reinforces the existing literature using Big Data, presents alternative ways to identify CIDs, and illustrates diversity within the image dimensions.
•City image dimensions (CIDs) were identified using 222,000 photos of 222 cities.•Google Cloud Vision AI and latent Dirichlet allocation (LDA) were used.•Universal CIDs are cityscape, landscape, transport, architecture, recreation.•Individual CIDs of Hong Kong, Rome, Montreal, Baghdad were identified.•Exploration of regional CIDs and CIDs based on quality of life.
The study re-examines the tourist-gaze concept using Bangkok, a Fareast tourism destination as a context of investigation. The study enriches the tourist-gaze analysis and overcomes the ...methodological challenge by applying multimethod approaches. Latent Dirichlet Allocation and spatial analysis are complemented by Visual Content Analysis to delve into the foci of gazes of the international tourist groups. The study reveals similarities and differences of the gazes and the spatial characteristics when the tourists visited Bangkok. The potential gaze dichotomies are also discussed. The findings enrich former debates on the influence of Western-centric imagery, and the photo representations of Bangkok in completing hermeneutic circle.
For intermediate and high risk prostate cancer, both the prostate gland and seminal vesicles are included in the clinical target volume. Internal motion patterns of these two organs vary, presenting ...a challenge for adaptive treatment. Adaptive techniques such as isocenter repositioning and soft tissue alignment are effective when tumor volumes only exhibit translational shift, while direct re-optimization of the intensity-modulated radiation therapy (IMRT) plan maybe more desirable when extreme deformation or differential positioning changes of the organs occur. Currently, direct re-optimization of the IMRT plan using beamlet (or fluence map) has not been reported. In this study, we report a novel on-line re-optimization technique that can accomplish plan adjustment on-line. Deformable image registration is used to provide position variation information on each voxel along the three dimensions. The original planned dose distribution is used as the 'goal' dose distribution for adaptation and to ensure planning quality. Fluence maps are re-optimized via linear programming, and a plan solution can be achieved within 2 min. The feasibility of this technique is demonstrated with a clinical case with large deformation. Such on-line ART process can be highly valuable with hypo-fractionated prostate IMRT treatment.
Well-known tourist attractions receive reviews in various languages. Google Maps has a reviewing system with translation service. The translated reviews may not perfectly represent the original ...review, but may still capture valuable information. This paper studies the effectiveness of nonusing translated reviews in sentiment analysis for tourist attractions in Phuket, Thailand. Reviews for beaches in Phuket and nearby islands are collected. The reviews are separated into two groups, reviews originally written in English and non-English. Rather than discarding non-English reviews or constructing models for each origin language, we experiment with different ways of utilizing the non-English reviews by translating them into English. Results show that models constructed using the reviews originally written in English are also effective for translated reviews. This can significantly reduce the effort to understand the sentiment of reviews written in various languages by training only one model and applying it to text translated from any language.
A center-crossing recurrent neural network is one in which the null- (hyper) surfaces of each neuron intersect at their exact centers of symmetry, ensuring that each neuron's activation function is ...centered over the range of net inputs that it receives. We demonstrate that relative to a random initial population, seeding the initial population of an evolutionary search with center-crossing networks significantly improves both the frequency and the speed with which high-fitness oscillatory circuits evolve on a simple walking task. The improvement is especially striking at low mutation variances. Our results suggest that seeding with center-crossing networks may often be beneficial, since a wider range of dynamics is more likely to be easily accessible from a population of center-crossing networks than from a population of random networks.
The challenges of real-time Gamma Knife™ inverse planning are the large number of variables involved and the unknown search space a priori. With limited collimator sizes, shots have to be heavily ...overlapped to form a smooth prescription isodose line that conforms to the irregular target shape. Such overlaps greatly influence the total number of shots per plan, making pre-determination of the total number of shots impractical. However, this total number of shots usually defines the search space, a pre-requisite for most of the optimization methods. Since each shot only covers part of the target, a collection of shots in different locations and various collimator sizes selected makes up the global dose distribution that conforms to the target. Hence, planning or placing these shots is a combinatorial optimization process that is computationally expensive by nature. We have previously developed a theory of shot placement and optimization based on skeletonization. The real-time inverse planning process, reported in this paper, is an expansion and the clinical implementation of this theory. The complete planning process consists of two steps. The first step is to determine an optimal number of shots including locations and sizes and to assign initial collimator size to each of the shots. The second step is to fine-tune the weights using a linear-programming technique. The objective function is to minimize the total dose to the target boundary (i.e., maximize the dose conformity). Results of an ellipsoid test target and ten clinical cases are presented. The clinical cases are also compared with physician’s manual plans. The target coverage is more than 99% for manual plans and 97% for all the inverse plans. The RTOG PITV conformity indices for the manual plans are between 1.16 and 3.46, compared to 1.36 to 2.4 for the inverse plans. All the inverse plans are generated in less than 2 min, making real-time inverse planning a reality.