Primary Dataset Research Articles

Asymmetric response of gross primary productivity to precipitation changes over China’s karst areas

China’s karst areas play an important role in regulating the global carbon dynamic and mitigating atmospheric CO2 concentration. Yet, extreme drought events have occurred frequently in recent years over China’s karst areas, which have a negative effect on the gross primary productivity (GPP) in the region. It is unclear whether gross primary productivity gains in the wet years can compensate for its losses in the dry years over the karst areas. In this study, we selected the asymmetry indices to assess the potential asymmetric response of gross primary productivity to the change in precipitation over China’s karst areas during the 2003 to 2018 period, using the gross primary productivity and precipitation dataset. Our results show that the gross primary productivity exhibits a positive asymmetry in response to precipitation changes, namely, the gains caused by increased precipitation in the wet years overcompensate the losses caused by decreased precipitation in the dry years. In addition, the gross primary productivity asymmetry shows an increasing trend over the study period despite extreme drought events occurring frequently, which is due to the response of gross primary productivity to drought has significantly decreased over the study period. For each biome, grasslands show the highest positive gross primary productivity asymmetry, indicating that the grassland biomes have a stronger capacity to utilize the increased precipitation during the wet years to increase gross primary productivity compared to other biomes. Furthermore, the gross primary productivity asymmetry over China’s karst areas can be effected by the precipitation asymmetry as well as mean annual precipitation. Our results will contribute to our knowledge of the response of gross primary productivity to precipitation changes in the karst areas of China.

An Artificial Intelligence based automated case-based reasoning (CBR) system for severity investigation and root-cause analysis of road accidents – Comparative analysis with the predictions of ChatGPT

Road accidents have been progressively causing havoc in our society and certain preventive measures must be taken to reduce or possibly eliminate road accidents. The derivative of a road accident ranges from a mild injury to casualty. This research work mainly focusses on developing a novel case-based reasoning system to investigate and troubleshoot the cause of road accidents on a war-foot basis. First, the dominant attributes contributing to the cause of road accidents are identified and finalized as 28. A unique road accident dataset is developed which comprises of 1028 data collected from web resources, popular news magazines and extended further to large scale database of one-million cases by biased random number simulation. Each attribute is given a severity weightage of 1,2 and 3 for computing the net weighted score for a case in the database. Also, non-weighted scores are computed by introduction of a primary number dataset to maintain the uniqueness of the score which is further used for similarity analytics. Now, an accident news is randomly selected, and Rapid automatic keyword extraction (RAKE) schema is used as Natural language processor (NLP) for extracting the dominant keywords from the news articles. The extracted keywords are compared and further mapped into a factor-matrix comprising 28 attributes causing road accidents. Further, similarity analytics is performed to evaluate the severity scores and comparison of new cases. The system demonstrated high retrieval accuracy with all road accident cases collected from real world scenarios. This research has great prospects on troubleshooting road accident cases effectively and provides instant promising troubleshooting measures to prevent such accidents in the future. Also, the proposed framework might be useful for intelligent decision-making systems and automated driving systems. Based on the final outlook, a comprehensive framework for national road safety could be developed and passed as a valid law for implementation. Finally, the forecasted results of the proposed algorithm are compared with the predictions of Chat GPT program.

Revealing AGNs through TESS variability

ABSTRACT We used Transiting Exoplanet Survey Satellite (TESS) data to identify 29 candidate active galactic nuclei (AGNs) through their optical variability. The high-cadence, high-precision TESS light curves present an opportunity for the identification of AGNs, including those not selected through other methods. Of the candidates, we found that 18 have either previously been identified as AGNs in the literature or could have been selected based on emission-line diagnostics, mid-IR colours, or X-ray luminosity. AGNs in low-mass galaxies offer a unique window into supermassive black hole and galaxy co-evolution and 8 of the 29 candidates have estimated black hole masses ≲ 106 M⊙. The low-mass galaxies NGC 4395 and NGC 4449 are two of our five ‘high-confidence’ candidates. Since our initial sample largely draws from just nine TESS sectors, we expect to identify at least ∼45 more candidates in the TESS primary and extended mission data sets, of which ∼60 per cent will be new AGNs and ∼20 per cent will be in low-mass galaxies.

Evaluation of energy generation potential from municipal solid waste in the North-West province, South Africa

Municipal Solid Waste (MSW) management is rapidly becoming a severe environmental problem worldwide. Developing countries, especially African cities, are the most affected due to inadequate resources to cope with increasing magnitude and complexity of the waste generated as well as the scarcity of land for disposal. As such, strategies which include waste- to-energy (WtE) generation to recover the potent energy from municipal waste could be a better option. This study sought to determine the sustainability of WtE projects for energy access to off-grid residents in the North-West province, South Africa. The study used a quantitative research design coupled with field observations and measurement of elements of the waste chain to generate primary data sets. The information was supplemented by secondary datasets on waste information and waste management at local municipalities. Results revealed that some of the classes of waste have the optimum calorific values and moisture content for WtE. The eligibility of a waste class to be used in WtE generation projects is dependent on the quantities generated. The results also indicate that using paper as fuel in the 240 tonnes/day WtE technology would cover more days of operation than plastics and rubber. Based on the 2020 estimated waste quantities, paper would last 234 days of generating about 6,944 GWh while plastics would last 177 days with energy output of 5, 207 GWh. Waste quantities generated in the North-West province could contribute to sustainable energy access to the off-grid informal settlement residents and advance waste management options through WtE. This study contributes to the literature on renewable energy and waste management in the context of green energy in South Africa.

CUSTOMER SEGMENTATION IN SUPERMARKET USING CLUSTER ANALYSIS

The research aims to understand the types of customer groups shopping in supermarkets, the segmentation of customers is based on the variables which impact the decision-making of their preference towards supermarkets. Cluster analysis is performed on a primary dataset collected in the form of a questionnaire. The questionnaire consists of the variables that can influence customers' preference towards supermarkets. In total 86 respondents did participate in data collection. The analysis of the data was piloted with hierarchal cluster analysis followed by K- Means cluster analysis. The finding of the analysis suggests that the customers can be segmented into 3 groups. The decisions related to offers, programs, priorities, promotion, place and other activities can be planned according to the target segment and market need. Keywords: Cluster Analysis, Supermarket, Consumer preference.

ConFruit: Effective Fruit Classification Using CNN Algorithm

Fruit is one type of food containing nutrients, vitamins, and minerals that are generally very good for daily consumption. However, various fruit choices make consumers confused about choosing and buying fruit. Many papers have proposed fruit classification to deal with this problem in recent years. Therefore, this study offers a new recommendation model using type to dissect fruit so that buyers can more easily recognize fruit. We collected the primary dataset from Cagle to 3000 fruit images. Based on experiments, our research achieved good accuracy results using the CNN algorithm to classify fruit so that consumers can distinguish between types of fruit. Experimentally demonstrated, we harvested the promised results with better accuracy and small losses than the general fruit classification study.

The Personal Health Applications of Machine Learning Techniques in the Internet of Behaviors

With the swift pace of the development of artificial intelligence (AI) in diverse spheres, the medical and healthcare fields are utilizing machine learning (ML) methodologies in numerous inventive ways. ML techniques have outstripped formerly state-of-the-art techniques in medical and healthcare practices, yielding faster and more precise outcomes. Healthcare practitioners are increasingly drawn to this technology in their initiatives relating to the Internet of Behavior (IoB). This area of research scrutinizes the rationales, approaches, and timing of human technology adoption, encompassing the domains of the Internet of Things (IoT), behavioral science, and edge analytics. The significance of ML in medical and healthcare applications based on the IoB stems from its ability to analyze and interpret copious amounts of complex data instantly, providing innovative perspectives that can enhance healthcare outcomes and boost the efficiency of IoB-based medical and healthcare procedures and thus aid in diagnoses, treatment protocols, and clinical decision making. As a result of the inadequacy of thorough inquiry into the employment of ML-based approaches in the context of using IoB for healthcare applications, we conducted a study on this subject matter, introducing a novel taxonomy that underscores the need to employ each ML method distinctively. With this objective in mind, we have classified the cutting-edge ML solutions for IoB-based healthcare challenges into five categories, which are convolutional neural networks (CNNs), recurrent neural networks (RNNs), deep neural networks (DNNs), multilayer perceptions (MLPs), and hybrid methods. In order to delve deeper, we conducted a systematic literature review (SLR) that examined critical factors, such as the primary concept, benefits, drawbacks, simulation environment, and datasets. Subsequently, we highlighted pioneering studies on ML methodologies for IoB-based medical issues. Moreover, several challenges related to the implementation of ML in healthcare and medicine have been tackled, thereby gradually fostering further research endeavors that can enhance IoB-based health and medical studies. Our findings indicated that Tensorflow was the most commonly utilized simulation setting, accounting for 24% of the proposed methodologies by researchers. Additionally, accuracy was deemed to be the most crucial parameter in the majority of the examined papers.

A new in-hospital mortality prediction nomogram for intensive care unit patients with acute pancreatitis.

Acute pancreatitis (AP) is a prevalent inflammatory disease that can lead to severe abdominal pain and multiple organ failure, potentially resulting in pancreatic necrosis and persistent dysfunction. A nomogram prediction model was developed to accurately evaluate the prognosis and provide therapy guidance to AP patients. Retrospective data extraction was performed using MIMIC-IV, an open-source clinical database, to obtain 1344 AP patient records, of which the primary dataset included 1030 patients after the removal of repeated hospitalizations. The prediction of in-hospital mortality (IHM) used the least absolute shrinkage and selection operator (LASSO) regression model to optimize feature selection. A multivariate logistic regression analysis was used to build a prediction model incorporating the selected features, and the C-index, calibration plot, and decision curve analysis (DCA) were utilized to evaluate the discrimination, calibration, and clinical applicability of the prediction model. The nomogram utilized a combination of indicators, including the SAPS II score, RDW, MBP, RR, PTT, and fluid-electrolyte disorders. Impressively, the model exhibited a satisfactory diagnostic performance, with area under the curve values of 0.892 and 0.856 for the training cohort and internal validation, respectively. Moreover, the calibration plots and the Hosmer-Lemeshow goodness-of-fit (HL) test revealed a strong correlation between the predicted and actual outcomes (p = 0.73), further confirming the reliability of our model. Notably, the results of the decision curve analysis (DCA) highlighted the superiority of our model over previously described scoring methods in terms of net clinical benefit, solidifying its value in clinical applications. Our novel nomogram is a simple tool for accurately predicting IHM in ICU patients with AP. Treatment methods that enhance the factors involved in the model may contribute to increased in-hospital survival for these ICU patients.

The correlation between primary open-angle glaucoma (POAG) and gut microbiota: a pilot study towards predictive, preventive, and personalized medicine.

Glaucoma is the leading cause of irreversible blindness worldwide. Emerged evidence has shown that glaucoma is considered an immune system related disorder. The gut is the largest immune organ in the human body and the gut microbiota (GM) plays an irreversible role in maintaining immune homeostasis. But, how the GM influences glaucoma remains unrevealed. This study aimed at investigating the key molecules/pathways mediating the GM and the glaucoma to provide new biomarkers for future predictive, preventive, and personalized medicine. Datasets from the primary open-angle glaucoma (POAG) patients (GSE138125) and datasets for target genes of GM/GM metabolites were downloaded from a public database. For GSE138125, the differentially expressed genes (DEGs) between healthy and POAG samples were identified. And the online Venn diagram tool was used to obtain the DEGs from POAG related to GM. After which GM-related DEGs were analyzed by correlation analysis, pathway enrichment analysis, and protein-protein interaction (PPI) network analysis. Human trabecular meshwork cells were used for validation, and the mRNA level of hub genes was verified by quantitative real-time polymerase chain reaction (RT-qPCR) in the in vitro glaucoma model. A total of 16 GM-related DEGs in POAG were identified from the above 2 datasets (9 upregulated genes and 7 downregulated genes). Pathway enrichment analysis indicated that these genes are mostly enriched in immune regulation especially macrophages-related pathways. Then 6 hub genes were identified by PPI network analysis and construction of key modules. Finally, RT-qPCR confirmed that the expression of the hub genes in the in vitro glaucoma model was consistent with the results of bioinformatics analysis of the mRNA chip. This bioinformatic study elucidates NFKB1, IL18, KITLG, TLR9, FKBP2, and HDAC4 as hub genes for POAG and GM regulation. Immune response modulated by macrophages plays an important role in POAG and may be potential targets for future predictive, preventive, and personalized diagnosis and treatment. The online version contains supplementary material available at 10.1007/s13167-023-00336-2.

A Machine Learning-Based Model to Predict In-Hospital Mortality of Lung Cancer Patients: A Population-Based Study of 523,959 Cases.

Background: Stratify new lung cancer patients based on the risk of in-hospital mortality rate after diagnosis. Methods: 522,941 lung cancer cases with available data on the Surveillance, Epidemiology, and End Results (SEER) were analyzed for the predicted probability based on six fundamental variables including age, gender, tumor size, T, N, and AJCC stages. The patients were randomly assigned to the training (n = 115,145) and validation datasets (n = 13,017). The remaining cohort with missing values (n = 394,779) was then combined with the primary lung tumour datasets (n = 1018) from The Cancer Genome Atlas, Lung Adenocarcinoma and Lung Squamous Cell Carcinoma projects (TCGA-LUAD & TCGA-LUSC) for external validation and sensitivity analysis. Results: Receiver Operating Characteristic (ROC) analyses showed high discriminatory power in the training and internal validation cohorts (Area under the curve [AUC] of 0.78 (95%CI = 0.78-0.79) and 0.78 (95%CI = 0.77-0.79), respectively), whereas that of the model on external validation data was 0.759 (95%CI = 0.757-0.761). We developed a static nomogram, a web app, and a risk table based on a logistic regression model using algorithm-selected variables. Conclusions: Our model can stratify lung cancer patients into high- and low-risk of in-hospital mortality to assist clinical further planning.

A standard for sharing spatial transcriptomics data

Spatial transcriptomic technologies have the potential to reveal critical relationships between the function of genes and cells and their spatial organization. Here, we provide a sharing model for spatial transcriptomics data with the aim of establishing a set of primary data and metadata needed to reproduce analyses and facilitate computational methods development.

What drives activity of banking agents? Evidence from Rural India

This article investigates previously unexplored predictors of banking agents’ activity using a primary dataset of agents operational in a rural setting in India. We measure banking agents’ activity using their monthly transaction count and regress it on seven predictor categories. We provide evidence of the role of operational practices, agents’ commitment, resources, support, and market characteristics, in agents’ banking activity. We also identify specific factors that drive superior performance by the agents. The findings inform policy-making groups and implementing agencies of agent banking on agent selection criteria, best practices, and supervision mechanisms to ensure the agent banking model’s effectiveness.

THE IMPACT OF CLIMATIC FACTORS AND FOOD SECURITY ON DEMOGRAPHIC FERTILITY IN PUNJAB, PAKISTAN THROUGH GEOSPATIAL STRUCTURAL MODELLING (GSSM)

: Punjab is not only the most fertile province of Pakistan but also the most fertile demographically with 3.4 total fertility rate. The fertility rate is an important demographic phenomenon that has significant implications for sustainable development, particularly in agricultural regions such as Punjab, Pakistan. In recent years, scholars and policymakers have increasingly focused on the link between fertility rate and soil fertility. As soil fertility is an important determinant of agricultural productivity and food security. This research article has explored the relationship and impact of the main climatic factors like maximum and minimum temperature, precipitation and Gross Primary Productivity on total fertility rate of districts of Punjab, for the year 2021. Geospatial Structural modeling (GSSM) has been designed to estimate the influence of climatic factors directly on fertility and its indirect influence on fertility through gross primary productivity and child mortality rate on total fertility rate of districts of Punjab. All the said geospatial primary and secondary datasets are analyzed in ArcGIS environment and resultant map are generated. The results shows that there is the direct and significant impact of gross primary productivity and child mortality rate on the total fertility rate. While there is indirect impact of mean annual precipitation, mean annual maximum and minimum temperature on the total fertility rate. Hence, there is negative direct impact of said factors on total fertility rate of the districts of the Punjab

Urban space maintainability factors based on life cycle approach

PurposeThe maintainability of urban spaces has become critical with rapid urbanization to create an effective and safe environment for the increasing population. The absence of scientific studies exploring the factors that affect urban space maintainability (USM) has hindered the incorporation of maintainability aspects during the urban space planning and designing stages. This paper aims to establish critical factors for USM.Design/methodology/approachQualitative content analysis is performed under an abductive approach to developing USM factors. A bibliometric search is conducted using databases including Scopus Elsevier, Emerald Insight, Science Direct, IEEE XPLORE and the American Society of Civil Engineers. The selected primary data set comprises journal papers on USM published after 2000. Seventy-three journal articles are selected through a comprehensive screening procedure and subjected to further analysis. The literature findings are processed via a software-assisted systematic coding and visualizing of the key data using NVivo 12 software. The coded USM factors are validated based on experts’ consensus statements by conducting an expert focus group discussion.FindingsTwelve critical factors are established for USM; they include six design stage-related factors, one construction stage-related factor and five operational stage-related factors.Research limitations/implicationsEstablished USM factors give an insight into the main focus areas when incorporating maintainability into urban spaces.Originality/valueThe authors establish a set of maintainability factors for urban spaces based on the life cycle stages. USM factors such as vegetation management, interdepartmental coordination and work zone safety draw attention to context-specific aspects of USM.

Longitudinal assessment of demographic representativeness in the Medical Imaging and Data Resource Center open data commons.

The Medical Imaging and Data Resource Center (MIDRC) open data commons was launched to accelerate the development of artificial intelligence (AI) algorithms to help address the COVID-19 pandemic. The purpose of this study was to quantify longitudinal representativeness of the demographic characteristics of the primary MIDRC dataset compared to the United States general population (US Census) and COVID-19 positive case counts from the Centers for Disease Control and Prevention (CDC). The Jensen-Shannon distance (JSD), a measure of similarity of two distributions, was used to longitudinally measure the representativeness of the distribution of (1)all unique patients in the MIDRC data to the 2020 US Census and (2)all unique COVID-19 positive patients in the MIDRC data to the case counts reported by the CDC. The distributions were evaluated in the demographic categories of age at index, sex, race, ethnicity, and the combination of race and ethnicity. Representativeness of the MIDRC data by ethnicity and the combination of race and ethnicity was impacted by the percentage of CDC case counts for which this was not reported. The distributions by sex and race have retained their level of representativeness over time. The representativeness of the open medical imaging datasets in the curated public data commons at MIDRC has evolved over time as the number of contributing institutions and overall number of subjects have grown. The use of metrics, such as the JSD support measurement of representativeness, is one step needed for fair and generalizable AI algorithm development.

Completeness of endocrine therapy information in the Primary Care Prescription Database (PCPD) and secondary care treatment datasets: A national population-based cohort study using routine healthcare data

BackgroundEndocrine therapy (ET) is a widely used treatment for breast cancer. In the UK, use is typically initiated in secondary care, with subsequent treatment in primary care. Evaluating use of ET depends on data sources containing accurate and complete information. This study aimed to evaluate the completeness and consistency of ET recorded in primary and secondary care data (SCD) and determine the value of combining data sources in describing use of ET. MethodsThis cohort study included women (50 + years) diagnosed with hormone receptor-positive invasive breast cancer in England, April-2015 to December-2019. Concordance of ET recorded in SCD and the Primary Care Prescription Database (PCPD) was evaluated. Factors associated with recording of ET in each setting were assessed using statistical models. ResultsOverall 110,529 women were included. 94% had ET recorded in either SCD or PCPD. ET captured in SCD varied from 3% (in Systemic Anti-Cancer Therapy data) to 52% (in the Cancer Outcomes and Services Dataset; COSD). By contrast, 93% of patients had an ET prescription in PCPD. Among patients with ET recorded, this was not captured in COSD for 45%. Capture in COSD was lowest for younger women, those with no comorbidity/frailty, with lower stage or HER2-positive disease, or with other treatments recorded. Overall concordance between COSD and PCPD was 57%, but varied substantially across NHS trusts (lowest decile≤28%; highest decile≥86%). Among women with ET recorded in both settings, the earliest record was in COSD for 97%; 59% of initial ET prescriptions recorded in COSD were not captured in PCPD. Combining PCPD and COSD data enabled estimation of ET duration. ConclusionsPCPD is vital for understanding the use of ET within this population. Completeness of SCD could be improved by ensuring information on first ET prescription is recorded. PCPD (linked to SCD) is a valuable resource for examining patterns of care for patients with cancer, including treatment duration and adherence.

Reassigning kerbside parking to cycling lanes: Is Robin Hood planning good for business?

ABSTRACT In built-up areas, creating segregated bicycle lanes involves trade-offs with other modes, such as repurposing existing kerbside parking. Can parking be eliminated without financially impacting local businesses? This study answers this question empirically, employing as a case study Boundary Street – the high street of a gentrifying inner-city suburb in Brisbane, Australia. It seeks to elucidate (1) the current use of on-street parking along Boundary Street; (2) the necessity for driving to Boundary Street, based on distances travelled; and (3) Boundary Street patrons’ spending habits and visitation patterns. Three sets of primary data are used, including a public forum, an intercept survey, and a series of systematic field observations. The results show that most visitors reach the area on foot; parking spaces are generally available along the street, contrary to local perceptions. Nearly a third of private car users travel less than 3 km to Boundary Street and more than half travel less than 10 km (in other words, they come from the inner city). Therefore, they do not need to travel by car – if distance was the only consideration. Based on these findings, it is concluded that reassigning at least one row of kerbside parking to cycling along Boundary Street would be appropriate.

The Use of K-Nearest Neighbour Approach for Crime Rate Prediction in Public Universities: Case Study Adekunle Ajasin University, Akungba–Akoko (AAUA), Ondo State, Nigeria

The prediction of crime rates has been a challenging problem for law enforcement agencies and policymakers. The main challenge in developing a predictive model for crime rate is the unavailability and quality of data. Crime data is often incomplete, inconsistent, and biased, which has always affected the accuracy of the scholars’ designed model. For example, the use of traditional methods such as manual crime analysis and statistical analysis are often limited by the volume and complexity of the data. The need for a more accurate and efficient predictive model that can handle large and complex datasets is increasing. This research developed a predictive model using the KNN algorithm that is into accurate prediction of crime rates and helps policymakers to make better decisions. The KNN algorithm was implemented using Python programming language with scikit-learn libraries. The primary dataset was collected from security units of Adekunle Ajasin University, Akungba Akoko and Akungba police station. The Algorithm was implemented on a Laboratory Local Area Network at the Information and Communication Technology Application Centre (ICTAC), Adekunle Ajasin University, Akungba Akoko. The performance evaluation was based on precision, recall, and F1 score. The results show that the KNN algorithm can accurately predict crime rates with an average accuracy of 88%. The model demonstrates that the KNN algorithm can be a useful tool for predicting crime rates. In addition, the model is hoped to enhance crime prevention strategies, reduce crime rates, and ultimately improve the quality of life in university communities. Keywords: Crime, Crime Rate, Prediction, K-Nearest Neighbour, Public Universities. Akinwumi, D.A. (2023): The Use of K-Nearest Neighbour Approach for Crime Rate Prediction in Public Universities: Case Study Adekunle Ajasin University, Akungba–Akoko (AAUA), Ondo State, Nigeria. Journal of Advances in Mathematical & Computational Science. Vol. 11, No. 2. Pp 11-24. dx.doi.org/10.22624/AIMS/MATHS/V11N2P1. Available online at www.isteams.net/mathematics-computationaljournal.

Video Object Segmentation Using Multi-Scale Attention-Based Siamese Network

Video target segmentation is a fundamental problem in computer vision that aims to segment targets from a background by learning their appearance information and movement information. In this study, a video target segmentation network based on the Siamese structure was proposed. This network has two inputs: the current video frame, used as the main input, and the adjacent frame, used as the auxiliary input. The processing modules for the inputs use the same structure, optimization strategy, and encoder weights. The input is encoded to obtain features with different resolutions, from which good target appearance features can be obtained. After processing using the encoding layer, the motion features of the target are learned using a multi-scale feature fusion decoder based on an attention mechanism. The final predicted segmentation results were calculated from a layer of decoded features. The video object segmentation framework proposed in this study achieved optimal results on CDNet2014 and FBMS-3D, with scores of 78.36 and 86.71, respectively. It outperformed the second-ranked method by 4.3 on the CDNet2014 dataset and by 0.77 on the FBMS-3D dataset. Suboptimal results were achieved on the video primary target segmentation datasets SegTrackV2 and DAVIS2016, with scores of 60.57 and 81.08, respectively.

Classification of head and neck cancer from PET images using convolutional neural networks

The aim of this study was to develop a convolutional neural network (CNN) for classifying positron emission tomography (PET) images of patients with and without head and neck squamous cell carcinoma (HNSCC) and other types of head and neck cancer. A PET/magnetic resonance imaging scan with 18F-fluorodeoxyglucose (18F-FDG) was performed for 200 head and neck cancer patients, 182 of which were diagnosed with HNSCC, and the location of cancer tumors was marked to the images with a binary mask by a medical doctor. The models were trained and tested with five-fold cross-validation with the primary data set of 1990 2D images obtained by dividing the original 3D images of 178 HNSCC patients into transaxial slices and with an additional test set with 238 images from the patients with head and neck cancer other than HNSCC. A shallow and a deep CNN were built by using the U-Net architecture for classifying the data into two groups based on whether an image contains cancer or not. The impact of data augmentation on the performance of the two CNNs was also considered. According to our results, the best model for this task in terms of area under receiver operator characteristic curve (AUC) is a deep augmented model with a median AUC of 85.1%. The four models had highest sensitivity for HNSCC tumors on the root of the tongue (median sensitivities of 83.3–97.7%), in fossa piriformis (80.2–93.3%), and in the oral cavity (70.4–81.7%). Despite the fact that the models were trained with only HNSCC data, they had also very good sensitivity for detecting follicular and papillary carcinoma of thyroid gland and mucoepidermoid carcinoma of the parotid gland (91.7–100%).