Small Database Research Articles

Succession Planning and Business Continuity of Family-Owned Business in Lagos State, Nigeria

This study investigates the effects of succession planning on the business continuity of family- owned businesses in Lagos state, Nigeria. The study adopted a cross-sectional survey design and computed a sample size of three hundred and sixty-eight (368) from The Nigerian Association of Small and Medium Enterprises NASME database of registered family businesses of eight-thousand three hundred and ninety- six registered family-owned businesses in Lagos state using the Cochran sample size formula. The senior staff and owners of family-owned businesses in Lagos state were purposively selected to fill out the structured questionnaires of the study. The questionnaires were adapted from previous studies and validated via a pilot study conducted in the Oluyole industrial area of Ibadan southwest, Oyo state, Nigeria. The study adopted SPSS version 25 for the descriptive statistics and Smart PLS version 4.0 for the inferential statistics to analyse the data. The study's findings revealed that succession planning influences 30.2% of business continuity, while the remaining 69.8% can be explained by the other exogenous variables different from business continuity. The study concludes that succession planning positively influences the business continuity of family-owned businesses in Lagos State, Nigeria.

An empirical method for estimating TBM penetration rate using tunnelling specific energy

Establishing a suitable tunnel boring machine (TBM) performance estimation model is benefit to the machine type selection, project scheduling and budgeting. Although various models have been developed to predict TBM performance, most of these models use a relatively small specific database, which is difficult to be applied to engineering practice. In this study, some equations for estimating TBM penetration rate (PR) applicable to different geological conditions are proposed. Firstly, a TBM construction performance database containing 159 tunnel sections is built according to four different tunnel projects originating from China, which covers rock mass properties, TBM specification parameters and performance information. By eliminating the effects of some TBM specification and operating parameters, a normalized energy index is proposed for assessing the rock mass boreability. On this basis, a series of statistical analysis, i.e. linear multiple regression (LMR) and non-linear multiple regression (NLMR) are used to develop some empirical equations for estimating the PR by employing rock mass properties and TBM specification parameters, and more than 84% of the PR can be accurately predicted. Finally, to further evaluate the prediction performance of these equations, they are compared and discussed with fifteen published typical models, and the results proved that the models developed in this study are more reasonable and reliable. It should be noted that the input parameters of these equations are readily available at the project planning stage, thus they are very simple and practical for designers and constructors.

Research on the method of eliminating differences between small-sample databases based on cloud computing

Abstract As computer technology continues to advance, more and more people are using databases, leading to variations in data backup and transfer between databases. This paper proposes a small sample inter-database discrepancy data elimination method based on cloud computing architecture as a way to solve the problem of data synchronization discrepancy between different databases. The semantic features of the data in the database are represented in the form of a directed graph, and the semantic Gaussian marginalized data fusion system is constructed by combining the rectangular window function of Gaussian marginalization to realize the fusion filtering processing of discrepant data. Then, the particle swarm discriminant tree algorithm is used to extract the features of the difference data between the small sample databases, and the KL transform is used to compress the difference data to improve its confidence level. The rough weighted average single dependency method is introduced to detect and identify the difference data between small sample databases and combined with the artificial intelligence algorithm to construct the principal component feature set of the difference data in the small sample databases, thus realizing the elimination of the difference data between small sample databases. When the proportion of difference data is increased from 0.05% to 1.00%, the leakage alarm rate and false alarm rate of this paper’s method for the difference data between small sample databases are 0.113% and 0.099%, respectively. When eliminating the inter-database discrepancy data, its time consumption is between 0.06μs and 0.3μs, and the average value of the removal rate of discrepancy data can reach 95.54%. Small sample databases that utilize cloud computing technology can utilize a variety of differential data elimination algorithms to ensure high-quality migration and synchronized backup of inter-database data.

Facial Beauty Prediction Based on Vision Transformer

Facial beauty analysis is a crucial subject in human culture among researchers across various applications. Recent studies have utilized multidisciplinary approaches to examine the relationship between facial traits, age, emotions, and other factors. Facial beauty prediction is a significant visual recognition challenge that evaluates facial attractiveness for human perception. This task demands considerable effort due to the novelty of the field and the limited resources available, including a small database for facial beauty prediction. In this context, a deep learning method has recently shown remarkable capabilities in predicting facial beauty. Additionally, vision Transformers have recently been introduced as novel deep learning approaches and have shown strong performance in various applications. The key issue is that the vision transformer performs significantly worse than ResNet when trained on a small ImageNet database. In this paper, we propose to address the challenges of predicting facial beauty by utilizing vision transformers instead of relying on feature extraction based on Convolutional Neural Networks, which are commonly used in traditional methods. Moreover, we define and optimize a set of hyperparameters according to the SCUT-FBP5500 benchmark dataset. The model achieves a Pearson coefficient of 0.9534. Experimental results indicated that using this proposed network leads to better predicting facial beauty closer to human evaluation than conventional technology that provides facial beauty assessment.

On machine learning analysis of atomic force microscopy images for image classification, sample surface recognition.

Atomic force microscopy (AFM or SPM) imaging is one of the best matches with machine learning (ML) analysis among microscopy techniques. The digital format of AFM images allows for direct utilization in ML algorithms without the need for additional processing. Additionally, AFM enables the simultaneous imaging of distributions of over a dozen different physicochemical properties of sample surfaces, a process known as multidimensional imaging. While this wealth of information can be challenging to analyze using traditional methods, ML provides a seamless approach to this task. However, the relatively slow speed of AFM imaging poses a challenge in applying deep learning methods broadly used in image recognition. This prospective is focused on ML recognition/classification when using a relatively small number of AFM images, aka small database. We discuss ML methods other than popular deep-learning neural networks. The described approach has already been successfully used to analyze and classify the surfaces of biological cells. It can be applied to recognize medical images, specific material processing, in forensic studies, even to identify the authenticity of arts. A general template for ML analysis specific to AFM is suggested, with a specific example of the identification of cell phenotype. Special attention is given to the analysis of the statistical significance of the obtained results, an important feature that is often overlooked in papers dealing with machine learning. A simple method for finding statistical significance is also described.

Entrepreneurial Orientation and Customer Satisfaction of Family-owned Business in Lagos State, Nigeria

This study investigates the effects of entrepreneurial orientation on customer satisfaction of family-owned businesses in Lagos State, Nigeria. The study adopted a cross-sectional survey design and computed a sample size of three hundred and sixty-eight (368) from The Nigerian Association of Small and Medium Enterprises NASME database of registered family businesses of eight thousand three hundred and ninety-six registered family-owned businesses in Lagos State using the Cochran sample size formula. The senior staff and owners of family-owned businesses in Lagos State were purposively selected to fill out the structured questionnaires of the study. The questionnaires were adapted from previous studies and validated via a pilot study conducted in the Oluyole industrial area of Ibadan southwest, Oyo State, Nigeria. The study adopted SPSS version 25 for the descriptive statistics and SmartPLS version 4.0 for the inferential statistics to analyse the data. The study's findings revealed that entrepreneurial orientation influences 31.6% of customer satisfaction, while the remaining 68.4% can be explained by the other exogenous variables different from customer satisfaction. The study concludes that entrepreneurial orientation positively influences the customer satisfaction of family-owned businesses in Lagos State, Nigeria.

Investigating the impact of human blood metabolites on the Sepsis development and progression: a study utilizing two-sample Mendelian randomization.

Existing data suggests a potential link between human blood metabolites and sepsis, yet the precise cause-and-effect relationship remains elusive. By using a two-sample Mendelian randomization (MR) analysis, this study aims to establish a causal link between human blood metabolites and sepsis. A two-sample MR analysis was employed to investigate the relationship between blood metabolites and sepsis. To assess the causal connection between sepsis and human blood metabolites, five different MR methods were employed, A variety of sensitivity analyses were conducted, including Cochrane's Q test, MR-Egger intercept test, MR-PRESSO and leave-one-out (LOO) analysis. In order to ensure the robustness of the causal association between exposure and outcome, the Bonferroni adjustment was employed. Additionally, we conducted analyses of the metabolic pathways of the identified metabolites using the Kyoto Encyclopedia of Genes and Genomes (KEGG) and the Small Molecule Pathway Database (SMPDB) database. The MR analysis revealed a total of 27 metabolites (16 known and 11 unknown) causally linked to the development and progression of sepsis. After applying the Bonferroni correction, 3-carboxy-4-methyl-5-propyl-2-furanpropanoate (CMPF) remained significant in relation to 28-day all-cause mortality in sepsis. By pathway enrichment analysis, we identified four significant metabolic pathways. Notably, the Alpha Linolenic Acid and Linoleic Acid metabolism pathway emerged as a pivotal contributor to the occurrence and progression of sepsis. This study provides preliminary evidence of causal associations between human blood metabolites and sepsis, as ascertained by MR analysis. The findings offer valuable insights into the pathogenesis of sepsis and may provide insight into preventive and therapeutic approaches.

Rational design of FXR agonists: a computational approach for NASH therapy.

Non-alcoholic fatty liver disease (NAFLD) is a hepatic manifestation of the metabolic syndrome, posing risks to cardiovascular and hepatic health worldwide. Non-alcoholic steatohepatitis (NASH) which is a severe form of NAFLD, has a global prevalence. Therapeutic targets for NASH include THR-β, GLP-1 receptor, PPARα/δ/γ, FGF21 analogs, and FXR, a bile acid nuclear receptor pivotal for regulating bile acid synthesis and excretion. Our study aims to design the non-steroidal FXR agonist for NASH treatment, as FXR's role in the regulation of bile acid processes, rendering it a promising drug target for NASH therapy. Utilizing tropifexor as a reference molecule, we generated a shape-based pharmacophore model with seven features, identifying key binding requirements within the FXR active site. Virtual screening using this model, coupled with molecular docking studies, helped pinpoint potential ligands from diverse small molecule databases. Further analysis via MM/GBSA revealed 12 molecules with binding affinities comparable to tropifexor. Among them, DB15416 exhibited the lowest binding free energy and superior docking scores. To assess its dynamic stability, we subjected DB15416 to molecular dynamics simulations, confirming its suitability as a FXR agonist. These findings suggest that DB15416 holds promise as a FXR agonist for NASH treatment, which can be evaluated by experimental studies.

PDBe CCDUtils: an RDKit-based toolkit for handling and analysing small molecules in the Protein Data Bank

While the Protein Data Bank (PDB) contains a wealth of structural information on ligands bound to macromolecules, their analysis can be challenging due to the large amount and diversity of data. Here, we present PDBe CCDUtils, a versatile toolkit for processing and analysing small molecules from the PDB in PDBx/mmCIF format. PDBe CCDUtils provides streamlined access to all the metadata for small molecules in the PDB and offers a set of convenient methods to compute various properties using RDKit, such as 2D depictions, 3D conformers, physicochemical properties, scaffolds, common fragments, and cross-references to small molecule databases using UniChem. The toolkit also provides methods for identifying all the covalently attached chemical components in a macromolecular structure and calculating similarity among small molecules. By providing a broad range of functionality, PDBe CCDUtils caters to the needs of researchers in cheminformatics, structural biology, bioinformatics and computational chemistry.Graphical

Lithium-ion battery state-of-health estimation: A self-supervised framework incorporating weak labels

The State-of-Health (SOH) estimation of Lithium-ion (Li-ion) batteries is critical for the safe and reliable operation of the batteries. Deep learning technologies are currently the popular methods for SOH estimation due to the advantages of no modeling and automatic feature extraction. However, existing methods require a large amount of annotated data to ensure model fitting, and the collection and labeling of battery aging data are time-consuming and laborious. Therefore, a self-supervised framework incorporating weak labels (SSF-WL) is proposed in this paper to obtain excellent estimation results on a small amount of annotated data. First, a novel data processing method based on the Gramian angular field, difference calculation, and raw data is proposed to enrich information and enhance features. Then, a five-layer Transformer encoder is constructed in SSF-WL for feature extraction. Finally, the model is pre-trained and fine-tuned on the proposed SSF-WL to obtain the estimated results of SOH. The proposed method is validated on the 124 commercial battery and Oxford databases. Experiments indicate that when using only 30% of the annotated training data, the Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) obtained by SSF-WL are 0.5219%/0.6085% lower than traditional supervised learning on the 124 commercial battery database, respectively. Moreover, the SSF-WL pre-trained model on a large unannotated database can be transferred to different types of batteries with a small annotated database and obtains on-par or better estimation results than the model trained from scratch.

PathBank 2.0-the pathway database for model organism metabolomics.

PathBank (https://pathbank.org) and its predecessor database, the Small Molecule Pathway Database (SMPDB), have been providing comprehensive metabolite pathway information for the metabolomics community since 2010. Over the past 14 years, these pathway databases have grown and evolved significantly to meet the needs of the metabolomics community and respond to continuing changes in computing technology. This year's update, PathBank 2.0, brings a number of important improvements and upgrades that should make the database more useful and more appealing to a larger cross-section of users. In particular, these improvements include: (i) a significant increase in the number of primary or canonical pathways (from 1720 to 6951); (ii) a massive increase in the total number of pathways (from 110234 to 605359); (iii) significant improvements to the quality of pathway diagrams and pathway descriptions; (iv) a strong emphasis on drug metabolism and drug mechanism pathways; (v) making most pathway images more slide-compatible and manuscript-compatible; (vi) adding tools to support better pathway filtering and selecting through a more complete pathway taxonomy; (vii) adding pathway analysis tools for visualizing and calculating pathway enrichment. Many other minor improvements and updates to the content, the interfaceand general performance of the PathBank website have also been made. Overall, we believe these upgrades and updates should greatly enhance PathBank's ease of use and its potential applications for interpreting metabolomics data.

Abstract 17760: Conversion of Dicom ECG Images to Tabular Format for Building Large Language Model in Diagnoses and Disease Progression of Cardiovascular Conditions

Introduction: DICOM Electrocardiography (ECG) images from individuals are routinely stored at the institutional PACS server, including normal and abnormal findings. Hypothesis: The study’s objective is to create a database of ECGs as per Standard Communication Protocol (SCP) and to build an accurate Large Language Model (LLM) across different cardiovascular conditions Methods: DICOM ECGs are retrieved, anonymized and labelled to convert the signals to (x, y) coordinates with the help of PyDicom libraries. The tabular format is then stored with lead definition and dimensions of duration and amplitude, metadata, clinical conditions, and textual diagnoses. Results: In the pilot study, 1308 ECGs from individuals with different age (<45 & >45 years), gender (male & female), binary clinical categories (normal & abnormal) and heart rate (<90 & >90 per min) are selected for analysis and modeling. The DICOM image signal reproducibility following conversion (Figure 1 & 2) shows Cross Correlation Percentage of 0.85 (0.76 - 0.94) (Figure 3). An eXtreme Gradient Boosting (XGBoost) model (6) was trained to predict above variables. The results (Figure 4) show AUC highest for predicting binary clinical categories (normal vs abnormal) at 0.81 (0.7 - 0.9) and lowest for predicting age (<45 & >45 years) at 0.6 (0.46 - 0.7). The accuracies in the leads also show similar trend. The limitations include reproducibility of results in Lead VIII, IX and XI, which are sub average and hence being retrained with optimizers. These are initial results with relatively smaller database where multiclass disease or condition classifications are not performed. Conclusions: To conclude, the methodology possesses opportunities to improve the model with Deep Learning and Entity Disambiguation Techniques (NLP) to build Large Language ECG Models for cardiovascular diagnoses and disease progression trajectories.

ECG-CL: A Comprehensive Electrocardiogram Interpretation Method Based on Continual Learning.

The value of Electrocardiogram (ECG) monitoring in early cardiovascular disease (CVD) detection is undeniable, especially with the aid of intelligent wearable devices. Despite this, the requirement for expert interpretation significantly limits public accessibility, underscoring the need for advanced diagnosis algorithms. Deep learning-based methods represent a leap beyond traditional rule-based algorithms, but they are not without challenges such as small databases, inefficient use of local and global ECG information, high memory requirements for deploying multiple models, and the absence of task-to-task knowledge transfer. In response to these challenges, we propose a multi-resolution model adept at integrating local morphological characteristics and global rhythm patterns seamlessly. We also introduce an innovative ECG continual learning (ECG-CL) approach based on parameter isolation, designed to enhance data usage effectiveness and facilitate inter-task knowledge transfer. Our experiments, conducted on four publicly available databases, provide evidence of our proposed continual learning method's ability to perform incremental learning across domains, classes, and tasks. The outcome showcases our method's capability in extracting pertinent morphological and rhythmic features from ECG segmentation, resulting in a substantial enhancement of classification accuracy. This research not only confirms the potential for developing comprehensive ECG interpretation algorithms based on single-lead ECGs but also fosters progress in intelligent wearable applications. By leveraging advanced diagnosis algorithms, we aspire to increase the accessibility of ECG monitoring, thereby contributing to early CVD detection and ultimately improving healthcare outcomes.

Irradiated mechanical properties predicted by a machine learning method with the Fourier-transform-based feature extension

ABSTRACT High-dimensional nonlinear relationships between the irradiated yield strength and its influencing factors, including doses, temperatures, and crystal structures, are difficult to explicitly characterize in the absence of a comprehensive database. In this study, we developed a machine learning method with the Fourier-transform-based feature extension, successfully constructing the prediction model of irradiated yield strength by a relatively small and sparse database of irradiated material properties. The analysis suggests that the proposed feature extension method improves the training performances of machine learning with small dataset. And the present model is accurate and feasible for predicting the irradiated yielding behaviors. Furthermore, we attempt the inverse machine learning model to determine material properties and irradiation conditions according to the desired yield strength. Since the parameter combinations commensurate with a fixed strength are diverse, the optimal model is helpful in reversely calculating and optimizing material performances. The data-driven machine learning method, which can detect the implicit correlations among numerous data, exhibits great prospects in investigating irradiated mechanical properties and exploring multiscale links in the nuclear material field. This work holds the promise for optimizing the design of in-pile structural components and can be further extended to other machine learning problems with the small dataset.

Identification of novel small molecule inhibitors for solute carrier SGLT1; a computational exploration

Diabetes results in substantial disabilities, diminished quality of life, and mortality that imposes a huge economic burden on societies and governments worldwide. Despite the absence of specific oral therapies at present, there exists an urgent requirement to develop a novel drug for the treatment of diabetes mellitus. The membrane protein sodium glucose co-transporters (SGLT1) present a captivating therapeutic target for diabetes, given its pivotal role in facilitating glucose absorption in the small intestine, offering immense promise for potential therapeutic intervention. In this connection, the present study is aimed at identifying potential inhibitors of SGLT1 from a small molecule database, including compounds from both natural as well as synthetic origins. A comprehensive approach was employed, by integrating homology modeling, ligand-based pharmacophore modeling, virtual screening, and molecular docking simulation. The process resulted in the identification of 16 new compounds, featuring similar attributes as observed for the documented actives. In a systematic screening procedure, five potential virtual hits were selected for simulation studies followed by subsequent binding free energy calculations, providing deeper insight into the time-dependent behavior of protein-ligand complexes in a dynamic state. In conclusion, our findings demonstrated that the identified compounds, particularly compounds 81 and 91, exhibit enhanced stability and favorable binding affinities with the target protein, marking them promising candidates for further investigations. Communicated by Ramaswamy H. Sarma

Robust crystal structure identification at extreme conditions using a density-independent spectral descriptor and supervised learning

The increased time- and length-scale of classical molecular dynamics simulations have led to raw data flows surpassing storage capacities, necessitating on-the-fly integration of structural analysis algorithms. As a result, algorithms must be computationally efficient, accurate, and stable at finite temperature to reliably extract the relevant features of the data at simulation time. In this work, we leverage spectral descriptors to encode local atomic environments and build crystal structure classification models. In addition to the classical way spectral descriptors are computed, i.e. over a fixed radius neighborhood sphere around a central atom, we propose an extension to make them independent from the material’s density. Models are trained on defect-free crystal structures with moderate thermal noise and elastic deformation, using the linear discriminant analysis (LDA) method for dimensionality reduction and logistic regression (LR) for subsequent classification. The proposed classification model is intentionally designed to be simple, incorporating only a limited number of parameters. This deliberate simplicity enables the model to be trained effectively even when working with small databases. Despite the limited training data, the model still demonstrates inherent transferability, making it applicable to a broader range of scenarios and datasets. The accuracy of our models in extreme conditions (high temperature, high density, large deformation) is compared to traditional algorithms from the literature, namely adaptive common neighbor analysis (a-CNA), polyhedral template matching (PTM) and diamond structure identification (IDS). Finally, we showcase two applications of our method: tracking a solid–solid BCC-to-HCP phase transformation in Zirconium at high pressure up to high temperature, and visualizing stress-induced dislocation loop expansion in single crystal FCC Aluminum containing a Frank–Read source, at high temperature.

Brain Tumor Detection and Classification using Deep Learning

Abstract: This paper explores the application of convolutional neural networks (CNNs) in diagnosing brain tumors, focusing on children and the elderly who are most vulnerable to this type of cancer. Brain tumors pose significant health risks due to uncontrolled cell growth within the skull, making accurate classification challenging. The study utilizes CNNs, a popular machine learning approach, to classify 3260 brain magnetic resonance images into glioma, meningioma, pituitary, and no tumor categories. Three algorithms, CNN, VGG, and Densenet, are employed for classification. Early detection is vital for effective treatment, and computer-aided diagnostics show promise in assisting clinicians. This work provides an overview of traditional computer-aided tumor diagnosis and highlights the advantages of using CNNs. It discusses image segmentation, classification techniques based on CNN, Densenet, and VGG16, and serves as a guide for future tumor detection research. Data augmentation and min-max normalization enhance tumor cell contrast. The dense CNN model accurately classifies a small image database, and Densenet demonstrates remarkable performance. Experimental results show CNN achieving 45% accuracy on testing data, VGG achieving 51%, and Densenet achieving an impressive 86.49% during training, with high precision and a good F1 score. Overall, this paper emphasizes the potential of CNNs for brain tumor diagnosis and contributes to the development of computeraided diagnostic systems. It advances the field of medical image analysis and paves the way for further research in tumor detection.

Database and Molecular Visualisation for Learning Media of Teaching Computational Chemistry/Pharmacy in Undergraduate Level

Teaching chemistry in the information technology era has made it easy to understand topics related to the structure and properties of a molecule. One of the courses that intensively utilises information technology as a learning media is computational chemistry. This study measured student-learning outcomes during computational courses using molecular visualization software and databases. The instruments of this research are the results/answers of final exams and the assignments of the students in the semesters of 2018/2019 and 2019/2020 in the chemistry and pharmacy departments. A hundred students enrolled in these courses per year. The study was conducted for two years. The score results of six assignments for practical and four essay questions tests were analyzed. Students were asked to visualize molecules using the requested application in practical test. The assessment shows that project scores with the topic of using databases and visualization, a 'good' score profile of the results the two years. The student has ‘good’ score on small molecule database and visualization. The students mastered the Avogadro application more than VMD. Students generally understand to get data (computer files) from the database and then the visual constructions requested by the questions.

Identification of novel scaffolds targeting SIRT3 through molecular modeling techniques for the treatment of Hepatocellular carcinoma

Hepatocellular carcinoma is one of the top causes of cancer-related death globally. SIRT3 belongs to the Sirtuin family of proteins, a collection of NAD+-dependent enzymes that play a role in controlling several cellular functions, including metabolism, aging, and stress response. SIRT3 expression has been discovered to be often downregulated in HCC tissues relative to normal liver tissues. Hence, SIRT3 may function as a tumor suppressor in HCC. In the present study, pharmacophore-based virtual screening of a small molecule database was performed initially, and then the screened hits were docked to the active site of SIRT3 to choose the best binding modes. One co-crystal ligand (PDB name: 1NQ) was utilized as a template to generate pharmacophore model query. A total of 0.2 million compounds from the VITAS-M Lab database were downloaded and prepared for virtual screening. Following database preparation, ligand-based virtual screening was performed using the pharmacophore query model generated in the previous phase. The compounds with the same pharmacophoric characteristics as the query at the same distance were screened. There were a total of 74 hits that matched the query model. These compounds were then docked to the SIRT3 using the standard precision protocol of the glide tool. To select hits with high binding affinities, a threshold of −8 kcal/mol was used. Based on the glide gscore, two hits were chosen. These two hits were selected to investigate the stability of the protein-ligand complex by molecular dynamics simulation. All of these findings indicate that the selected hit compounds C1 and C2 can serve as lead compounds in inhibiting the biological activity of SIRT3 requiring further detailed investigations. Communicated by Ramaswamy H. Sarma

What drives small and medium enterprises to establish and terminate banking relationships?

PurposeThis paper investigates the factors crucial for small and medium enterprises (SMEs) in establishing business relationships with banks in Pakistan.Design/methodology/approachTo investigate how SMEs select banking relationships using criteria, such as decision factors, decision-makers, and decision processes, a comprehensive literature review was used to classify SMEs' decision factors for bank selection. A survey questionnaire was distributed to 200 SMEs, randomly selected from the Small and Medium Enterprise Development Authority database in Pakistan. Probit/Tobit model is estimated to explain the behavior of SMEs.FindingsThe results reveal that SMEs consider a bank's Reputation, Price, and Location essential while establishing bank relationships. SMEs tend to terminate relationships with banks when the Price and Location of the bank are considered essential factors in the relationship with the banks. Price and Location are necessary for SMEs to reduce banking relationships. The SMEs also tend to reduce if they get attractive offers, or the SMEs are recommended to make a banking relationship. This study also provides intuitions for bank policymakers to design policies to retain SME customers and attract new business relationships.Practical implicationsThe research emphasizes the importance of competitive and transparent pricing strategies in designing products for SMEs. Banks must prioritize their Reputation and credibility to attract and retain relationships with SMEs.Originality/valueThe study attempts to provide evidence on the SME-Bank relationship focusing on the factors that are crucial for SMEs to decide while establishing business relationships with banks. Also, most of the related literature focuses on developed countries; this research adds to the literature on SMEs' behavior, particularly in a developing country's context.