Supervised Machine Learning Techniques Research Articles

Boosting and Comparing Performance of Machine Learning Classifiers with Meta-heuristic Techniques to Detect Code Smell

Background: Continuous modifications, suboptimal software design practices, and stringent project deadlines contribute to the proliferation of code smells. Detecting and refactoring these code smells are pivotal to maintaining complex and essential software systems. Neglecting them may lead to future software defects, rendering systems challenging to maintain, and eventually obsolete. Supervised machine learning techniques have emerged as valuable tools for classifying code smells without needing expert knowledge or fixed threshold values. Further enhancement of classifier performance can be achieved through effective feature selection techniques and the optimization of hyperparameter values. Aim: Performance measures of multiple machine learning classifiers are improved by fine tuning its hyperparameters using various type of meta-heuristic algorithms including swarm intelligent, physics, math, and bio-based etc. Their performance measures are compared to find the best meta-heuristic algorithm in the context of code smell detection and its impact is evaluated based on statistical tests. Method: This study employs sixteen contemporary and robust meta-heuristic algorithms to optimize the hyperparameters of two machine learning algorithms: Support Vector Machine (SVM) and k-nearest Neighbors (k-NN). The No Free Lunch theorem underscores that the success of an optimization algorithm in one application may not necessarily extend to others. Consequently, a rigorous comparative analysis of these algorithms is undertaken to identify the best-fit solutions for code smell detection. A diverse range of optimization algorithms, encompassing Arithmetic, Jellyfish Search, Flow Direction, Student Psychology Based, Pathfinder, Sine Cosine, Jaya, Crow Search, Dragonfly, Krill Herd, Multi-Verse, Symbiotic Organisms Search, Flower Pollination, Teaching Learning Based, Gravitational Search, and Biogeography-Based Optimization, have been implemented. Results: In the case of optimized SVM, the highest attained accuracy, AUC, and F-measure values are 98.75%, 100%, and 98.57%, respectively. Remarkably, significant increases in accuracy and AUC, reaching 32.22% and 45.11% respectively, are observed. For k-NN, the best accuracy, AUC, and F-measure values are all perfect at 100%, with noteworthy hikes in accuracy and ROC-AUC values, amounting to 43.89% and 40.83%, respectively. Conclusion: Optimized SVM exhibits exceptional performance with the Sine Cosine Optimization algorithm, while k-NN attains its peak performance with the Flower Optimization algorithm. Statistical analysis underscores the substantial impact of employing meta-heuristic algorithms for optimizing machine learning classifiers, enhancing their performance significantly. Optimized SVM excels in detecting the God Class, while optimized k-NN is particularly effective in identifying the Data Class. This innovative fusion automates the tuning process and elevates classifier performance, simultaneously addressing multiple longstanding challenges.

ABDpred: Prediction of active antimicrobial compounds using supervised machine learning techniques.

Discovery of new antibiotics is the need of the hour to treat infectious diseases. An ever-increasing repertoire of multidrug-resistant pathogens poses an imminent threat to human lives across the globe. However, the low success rate of the existing approaches and technologies for antibiotic discovery remains a major bottleneck. In silico methods like machine learning (ML) deem more promising to meet the above challenges compared with the conventional experimental approaches. The goal of this study was to create ML models that may be used to successfully predict new antimicrobial compounds. In this article, we employed eight different ML algorithms namely, extreme gradient boosting, random forest, gradient boosting classifier, deep neural network, support vector machine, multilayer perceptron, decision tree, and logistic regression. These models were trained using a dataset comprising 312 antibiotic drugs and a negative set of 936 non-antibiotic drugs in a five-fold cross validation approach. The top four ML classifiers (extreme gradient boosting, random forest, gradient boosting classifier and deep neural network) were able to achieve an accuracy of 80 per cent and above during the evaluation of testing and blind datasets. We aggregated the top performing four models through a soft-voting technique to develop an ensemble-based ML method and incorporated it into a freely accessible online prediction server named ABDpred ( http://clinicalmedicinessd.com.in/abdpred/ ).

A Hybrid Machine Learning Approach for Predicting Student Performance Using Multi-class Educational Datasets

Prediction of students’ academic performance has garnered considerable interest, with many institutions seek to enhance students’ performance and their quality of education. The integration of both unsupervised and supervised machine learning techniques has demonstrated significant efficacy in predicting student performance. This paper explores the application of different machine learning methods in predicting student academic performance. Initially, Principal Component Analysis (PCA) was utilised to reduce the dataset’s dimensionality, thereby improving its visualisation. Subsequently, K-Means clustering was employed to segregate students into distinct groups, reflective of their learning behaviors. Afterwards, the observed clusters were utilised for training classification models to address each student cluster individually. This approach was implemented in a case study involving an undergraduate science course at a North American University (NAU) and the Open University Learning Analytics Dataset (OULAD). Empirical findings indicate that the combined use of Feedforward Dense Network (FDN), Random Forest (RF), and Decision Tree (DT), specifically in their clustered forms, outperforms other classifiers in predicting student academic performance effectively.

Application of a machine learning model to maximize the success rate in day trade operations on the American Stock Exchange

Daytrading has been showing a growing popularity in the world due to easy access via technology, the possibility of additional earnings and a large increase in courses and several mentors available on social networks. This scenario causes many people to be unprepared to enter this market that has a high risk and that end up causing many people to lose their savings. Considering this situation, this study proposes the analysis of the data of a daytrade strategy, applying a machine learning model to help the investor make better decisions. Data from November 2020 to July 2023 was used within the US market based on the company [AMD]. The method used was the supervised machine learning technique known as the decision tree model, which seeks to identify the probability of event and non-event within the scenarios proposed in this work. The results were analyzed using the confusion matrix, gauging the accuracy in the training and test base, applying several decision tree models in order to find the best model and accuracy in the test base. In this sense, an improvement in the assertiveness rate was observed with the application of the supervised machine learning model based on a decision tree.

Racial Tropes in the Foreign Policy Bureaucracy: A Computational Text Analysis

AbstractHow do racial stereotypes affect perceptions in foreign policy? Race and racism as topics have long been marginalized in the study of international relations but are receiving renewed attention. In this article we assess the role of implicit racial bias in internal, originally classified assessments by the US foreign policy bureaucracy during the Cold War. We use a combination of dictionary-based and supervised machine learning techniques to identify the presence of four racial tropes in a unique corpus of intelligence documents: almost 5,000 President's Daily Briefs given to Kennedy, Johnson, Nixon, and Ford. We argue and find that entries about countries that the US deemed “racialized Others”—specifically, countries in the Global South, newly independent states, and some specific regional groupings—feature an especially large number of racial tropes. Entries about foreign developments in these places are more likely to feature interpretations that infantilize, invoke animal-based analogies, or imply irrationality or belligerence. This association holds even when accounting for the presence of conflict, the regime type of the country being analyzed, the invocation of leaders, and the topics being discussed. The article makes two primary contributions. First, it adds to the revival of attention to race but gives special emphasis to implicit racialized thinking and its appearance in bureaucratic settings. Second, we show the promise of new tools for identifying racial and other forms of implicit bias in foreign policy texts.

Application of Two Machine Learning Techniques for Predicting and Classifying the Type II Diabetes mellitus

The main objective of the study is to evaluate the prediction and the classification accuracy of two Supervised Machine Learning Techniques which are linear discriminant analysis (LDA) and logistic regression analysis (LRA) using real data of Type II diabetes. The classification accuracy for both models was determined by the classification accuracy rate. LRA and LDA correctly classified 78.70%and 80.00% of the Type II diabetes mellitus (diabetics and non-diabetics) respectively. The LRA has sensitivity and specificity was 64.38% and 85.35% respectively and the LDA had a sensitivity and specificity of 70.88% and 84.77% respectively. Both algorithms had a good overall classification rate. In terms of proper classification rate, the LDA model slightly outperformed the LRA approach. In general, the findings of this study revealed that the LRA model appears to be appropriate for prediction accuracy while the LDA model appears to be appropriate for classification procedures.

Pulp Particle Classification Based on Optical Fiber Analysis and Machine Learning Techniques

In the pulp and paper industry, pulp testing is typically a labor-intensive process performed on hand-made laboratory sheets. Online quality control by automated image analysis and machine learning (ML) could provide a consistent, fast and cost-efficient alternative. In this study, four different supervised ML techniques—Lasso regression, support vector machine (SVM), feed-forward neural networks (FFNN), and recurrent neural networks (RNN)—were applied to fiber data obtained from fiber suspension micrographs analyzed by two separate image analysis software. With the built-in software of a commercial fiber analyzer optimized for speed, the maximum accuracy of 81% was achieved using the FFNN algorithm with Yeo–Johnson preprocessing. With an in-house algorithm adapted for ML by an extended set of particle attributes, a maximum accuracy of 96% was achieved with Lasso regression. A parameter capturing the average intensity of the particle in the micrograph, only available from the latter software, has a particularly strong predictive capability. The high accuracy and sensitivity of the ML results indicate that such a strategy could be very useful for quality control of fiber dispersions.

Harnessing Small-Molecule Analyte Detection in Complex Media: Combining Molecularly Imprinted Polymers, Electrolytic Transistors, and Machine Learning.

Small-molecule analyte detection is key for improving quality of life, particularly in health monitoring through the early detection of diseases. However, detecting specific markers in complex multicomponent media using devices compatible with point-of-care (PoC) technologies is still a major challenge. Here, we introduce a novel approach that combines molecularly imprinted polymers (MIPs), electrolyte-gated transistors (EGTs) based on 2D materials, and machine learning (ML) to detect hippuric acid (HA) in artificial urine, being a critical marker for toluene intoxication, parasitic infections, and kidney and bowel inflammation. Reduced graphene oxide (rGO) was used as the sensory material and molecularly imprinted polymer (MIP) as supramolecular receptors. Employing supervised ML techniques based on symbolic regression and compressive sensing enabled us to comprehensively analyze the EGT transfer curves, eliminating the need for arbitrary signal selection and allowing a multivariate analysis during HA detection. The resulting device displayed simultaneously low operating voltages (<0.5 V), rapid response times (≤10 s), operation across a wide range of HA concentrations (from 0.05 to 200 nmol L-1), and a low limit of detection (LoD) of 39 pmol L-1. Thanks to the ML multivariate analysis, we achieved a 2.5-fold increase in the device sensitivity (1.007 μA/nmol L-1) with respect to the human data analysis (0.388 μA/nmol L-1). Our method represents a major advance in PoC technologies, by enabling the accurate determination of small-molecule markers in complex media via the combination of ML analysis, supramolecular analyte recognition, and electrolytic transistors.

Feature-based search space characterisation for data-driven adaptive operator selection

Combinatorial optimisation problems are known as unpredictable and challenging due to their nature and complexity. One way to reduce the unpredictability of such problems is to identify features and the characteristics that can be utilised to guide the search using domain-knowledge and act accordingly. Many problem solving algorithms use multiple complementary operators in patterns to handle such unpredictable cases. A well-characterised search space may help to evaluate the problem states better and select/apply a neighbourhood operator to generate more productive new problem states that allow for a smoother path to the final/optimum solutions. This applies to the algorithms that use multiple operators to solve problems. However, the remaining challenge is determining how to select an operator in an optimal way from the set of operators while taking the search space conditions into consideration. Recent research shows the success of adaptive operator selection to address this problem. However, efficiency and scalability issues persist in this regard. In addition, selecting the most representative features remains crucial in addressing problem complexity and inducing commonality for transferring experience across domains. This paper investigates if a problem can be represented by a number of features identified by landscape analysis, and whether an adaptive operator selection scheme can be constructed using Machine Learning (ML) techniques to address the efficiency and scalability problem. The proposed method determines the optimal categorisation by analysing the predictivity of a set of features using the most well-known supervised ML techniques. The identified set of features is then used to construct an adaptive operator selection scheme. The findings of the experiments demonstrate that supervised ML algorithms are highly effective when building adaptable operator selectors.

Exploring the resilience of supplementary cementitious materials-based concrete to elevated temperatures via modern computing techniques

Researchers are focused on the production of sustainable materials in order to reduce the negative environmental impact of conventional concrete. Utilizing industrial by-products as alternative cementitious materials in concrete is an appropriate strategy for promoting sustainability in construction. This research employed supervised machine learning techniques, including gradient boosting (GB), random forest (RF), and extreme gradient boosting (X-GB), in order to predict the compressive strength of sustainable concrete when exposed to high temperatures. SHapley Additive exPlanation (SHAP) analysis provided input component relevance. The comparison of these methodologies indicated that X-GB performed remarkably well and had a superior R2 value of 0.94 when compared to GB and RF. The results were further supported by a Taylor diagram, which demonstrated that the X-GB model best fitted the data, surpassing both the GB and RF models. This study's findings demonstrated the potential of machine learning, and more particularly X-GB, for making accurate predictions of compressive strength in high-temperature concrete applications. The results of SHAP analysis may be useful for selecting the proportions of raw materials in the mix design to achieve the desired properties.

Heuristics for online three-dimensional packing problems and algorithm selection framework for semi-online with full look-ahead

In online three-dimensional packing problems (3D-PPs), unlike offline problems, items arrive sequentially and require immediate packing decisions without any information about the quantities and sizes of the items to come. Heuristic methods are of great importance in solving online problems to find good solutions in a reasonable amount of time. However, the literature on heuristics for online problems is sparse. As our first contribution, we developed a pool of heuristics applicable to online 3D-PPs with complementary performance on different sets of instances. Computational results showed that in terms of the number of used bins, in all problem instances, at least one of our heuristics had a better or equal performance compared to existing heuristics in the literature. The developed heuristics are also fully applicable to an intermediate class between offline and online problems, referred to in this paper as a specific type of “semi-online with full look-ahead”, which has several practical applications. In this class, as in offline problems, complete information about all items is known in advance (i.e., full look-ahead); however, due to time or space constraints, as in online problems, items should be packed immediately in the order of their arrival. As our second contribution, we presented an algorithm selection framework, building on developed heuristics and utilizing prior information about items in this specific class of problems. We used supervised machine learning techniques to find the relationship between the features of problem instances and the performance of heuristics and to build a prediction model. The results indicate an 88% accuracy in predicting (identifying) the most promising heuristic(s) for solving any new instance from this class of problems.

A Supervised Machine Learning Algorithms: Applications, Challenges, and Recommendations

Machine Learning (ML) is an advanced technology that empowers systems to acquire knowledge autonomously, eliminating the need for explicit programming. The fundamental objective of the machine learning paradigm is to equip computers with the ability to learn independently without human intervention. In the literature, categorization in data mining has received a lot of traction, with applications ranging from health to astronomy and finance to textual classification. The three learning methodologies in machine learning are supervised, unsupervised, and semi-supervised. Humans must give the appropriate input and output and offer feedback on the prediction accuracy throughout the training phase for supervised algorithms. Unsupervised learning methods differ from supervised learning methods because they do not require any training. However, supervised learning methods are more accessible to implement than unsupervised learning methods. This study looks at supervised learning algorithms commonly employed in data classification. The strategies are evaluated based on their objective, methodology, benefits, and drawbacks. It is anticipated that readers will be able to understand the supervised machine learning techniques for data classification.

Uncovering the factors that affect earthquake insurance uptake using supervised machine learning

The escalating threat of natural disasters to public safety worldwide underlines the crucial role of effective environmental risk management tools, such as insurance. This is particularly evident in the case of earthquakes that occurred in Oklahoma between 2011 and 2020, which were linked to wastewater injection, underscoring the need for earthquake insurance. In this regard, from a survey of 812 respondents in Oklahoma, USA, we used supervised machine learning techniques (i.e., logit, ridge, least absolute shrinkage and selection operator (LASSO), decision tree, and random forest classifiers) to identify the factors that influence earthquake insurance uptake and to predict individuals who would acquire earthquake insurance. Our findings reveal that influential factors that affect earthquake insurance uptake include demographic factors such as older age, male gender, race, and ethnicity. These were found to significantly influence the decision to purchase earthquake insurance. Additionally, individuals residing in rental properties were less likely to purchase earthquake insurance, while longer residency in Oklahoma had a positive influence. Past experience of earthquakes was also found to positively influence the decision to purchase earthquake insurance. Both decision trees and random forests demonstrated good predictive capabilities for identifying earthquake insurance uptake. Notably, random forests exhibited higher precision and robustness, emerging as an encouraging choice for earthquake insurance modeling and other classification problems. Empirically, we highlight the importance of insurance as an environmental risk management tool and emphasize the need for awareness and education on earthquake insurance as well as the use of supervised machine learning algorithms for classification problems.

Sea level variability and modeling in the Gulf of Guinea using supervised machine learning

The rising sea levels due to climate change are a significant concern, particularly for vulnerable, low-lying coastal regions like the Gulf of Guinea (GoG). To effectively address this issue, it is crucial to gain a comprehensive understanding of historical sea level variability, and the influencing factors, and develop a reliable modeling system for future projections. This knowledge is essential for informed planning and mitigation strategies aimed at protecting coastal communities and ecosystems. This study presents a comprehensive analysis of mean sea level anomaly (MSLA) trends in the GoG between 1993 and 2020, covering three distinct periods (1993–2002, 2003–2012, and 2013–2020). It investigates the connections between interannual sea level variability and large-scale oceanic and atmospheric forcings. Furthermore, the study evaluates the performance of supervised machine learning techniques to optimize sea level modeling. The findings reveal a consistent rise in MSLA linear trends across the basin, particularly pronounced in the northern region, with a total linear trend of 88 mm over the entire period. The highest decadal trend (38.7 mm) emerged during 2013–2020, with the most substantial percentage increment (100%) occurring in 2003–2012. Spatial variation in decadal sea-level trends was influenced by subbasin physical forcings. Strong interannual signals in the spatial sea level distribution were identified, linked to large-scale oceanic and atmospheric phenomena. Seasonal variations in sea level trends are attributed to seasonal changes in the forcing factors. The evaluation of supervised learning modeling methods indicates that Random Forest Regression and Gradient Boosting Machines are the most accurate, reproducing interannual sea level patterns in the GoG with 97% and 96% accuracy. These models could be used to derive regional sea level projections via downscaling of climate models. These findings provide essential insights for effective coastal management and climate adaptation strategies in the GoG.

Customer Churn Prediction Using Supervised Machine Learning Technique Along With Recommendation.

Customer Churn Prediction Using Supervised Machine Learning Technique Along With Recommendation.

Developing a Saturation Height Function (SHF) for the classified Hydraulic Flow Unit, a case study from the Middle Miocene Carbonate reservoir in the Song Hong Basin, Vietnam

The Middle Miocene carbonate is the main gas bearing reservoir in the southern part of the Song Hong basin. Carbonate reservoirs present challenges for engineers and geologists to characterize because of their complexity due to depositional and diagenetic processes. The log based calculated water saturation Sw for this complex carbonate reservoir bears a lot of uncertainties due to its heterogeneity that strongly affects to m and n values in the water saturation equation. As the calculated Sw strongly affects the Hydrocarbon Initially In Place (HIIP) estimation, the production performance prediction, etc., the alternative method using the Saturation Height Function (SHF) based on the capillary pressure was introduced. The SHF building process uses the combination of the result of HFU classification and permeability prediction given by machine learning methods, which have become a very useful tool for complex reservoir characterization such as carbonate reservoir for decades. Our obtained results allowed us to subdivide the carbonate reservoir into 5 Hydraulic Flow Unit (HFU) by using unsupervised machine learning methods, and then the capillary pressure (Pc) curves were classified and a proper saturation-height function using Skelt Harrison equation was assigned to individual HFU after testing with common SHF equations. The HFU number for each individual point was predicted based on the log data measured along the borehole using the supervised machine learning techniques while the Sw was calculated using the chosen SHF model and the FWL defined on log curves and pressure data from RCI measurement. The Sw computed from the SHF model is reliable and can be used for water saturation estimation for the whole field for known FWL and predicted HFU numbers.

Environmental Consequences in the Neighbourhood of Rapid Unplanned Urbanisation in Bangalore City

The knowledge of landscape dynamics aids in evolving strategies for the prudent management of natural resources to sustain ecosystem services. The availability of spatiotemporal remote sensing data with advancements in artificial intelligence (AI) and machine learning (ML) algorithms has aided in assessing the ecological status in urban environments, markedly revealing complex patterns and interactions. The current communication presents landscape dynamics in the Bengaluru Urban district from 1973 to 2022 using a supervised machine learning technique based on the Random Forest algorithm with temporal Landsat data, which showed a 51.86% increase in the built-up area and a 26.28% decrease in the green cover. Rapid unplanned urbanization after globalization and the opening up of Indian markets (in Bengaluru city) has witnessed erosion in the natural surface (waterbodies and green cover) in the neighborhood, which has been impacting the health of the environment and people. Computation of fragmentation indices showed a decline of the native green cover by 177.2 sq. km. in the southern part of the district. Likely land use changes are predicted using the Cellular Automata Markov model considering the base case scenario. The analyses revealed a further possible increase in built-up to 1536.08 sq. km, a decrease in green cover by 14.32 sq. km by 2038, and the disappearance of water bodies, which highlights the need to mitigate the adverse impacts of land use changes through planned urbanization considering the environment and livelihood of local communities. The decline of heat sinks such as water bodies and green cover would contribute to an increase in the land surface temperature (LST), which would affect the microclimate of Bengaluru, highlighting the need to sustain ecosystem services to support the livelihood of local communities. Understanding the ecological significance of diverse habitat characteristics of the urban region and the prediction of likely changes in a high degree of spatial heterogeneity would assist the decision-makers in framing appropriate policies.

Improving the Performance of Arabic Information Retrieval Systems: The Issue of Resolving Word Sense Disambiguation

This study aimed at assessing the performance and efficacy of the retrieval information (IR) systems implemented in three widely used search engines (Google, Bing, and Yahoo), specifically with regard to the challenge of word sense disambiguation in Arabic texts. Such a challenge has been confirmed to negatively influence the retrieval of the most relevant documents. Therefore, we extended the paradigm of using computational methods and natural language processing (NLP) tools, primarily tailored for processing English texts, to explore morphosyntactic as well as lexical issues disturbing the accuracy of Arabic IR systems. Findings revealed striking disparities in the efficacy of IR systems integrated into these search engines, which can be attributed to four principal challenges: (a) the intricate morpho-syntactic structures inherent in Arabic; (b) the idiosyncratic orthographical system of the Arabic script; (c) the multifaceted semantic flexibility of certain lexical elements; and (d) the intriguing diaglossic nature of Arabic, allowing for the coexistence of multiple linguistic varieties within a single discourse situation. Drawing from these findings, a series of solutions rooted in supervised machine learning techniques, including clustering models and adaptations based on geographic locations, are proposed. Moreover, the study advocates for the capacity of search engines to interpret queries across all Arabic varieties, encompassing vernacular dialects. Furthermore, the importance of search engines accommodating queries irrespective of the specific language adopted by users is underscored. While the research primarily centers on Arabic, its implications resonate beyond this language alone. By applying computational methodologies originally designed for English to Arabic, the study not only addresses the challenges specific to Arabic IR systems but also contributes valuable insights that transcend linguistic boundaries. Through a comparative lens, issues like word sense disambiguation between Arabic and English are juxtaposed, extracting lessons that can inform advancements in information retrieval for both languages.

Predicting Abnormal Respiratory Patterns in Older Adults Using Supervised Machine Learning on Internet of Medical Things Respiratory Frequency Data

Wearable Internet of Medical Things (IoMT) technology, designed for non-invasive respiratory monitoring, has demonstrated considerable promise in the early detection of severe diseases. This paper introduces the application of supervised machine learning techniques to predict respiratory abnormalities through frequency data analysis. The principal aim is to identify respiratory-related health risks in older adults using data collected from non-invasive wearable devices. This article presents the development, assessment, and comparison of three machine learning models, underscoring their potential for accurately predicting respiratory-related health issues in older adults. The convergence of wearable IoMT technology and machine learning holds immense potential for proactive and personalized healthcare among older adults, ultimately enhancing their quality of life.

A multidisciplinary approach to facies evaluation at regional level using well log analysis, machine learning, and statistical methods

Geological facies evaluation is crucial for the exploration and development of hydrocarbon reservoirs. To achieve accurate predictions of litho-facies in wells, a multidisciplinary approach using well log analysis, machine learning, and statistical methods was proposed for the Lower Indus Basin. The study utilized five supervised machine learning techniques, including Random Forest (FR), Support Vector Machine (SVM), Artificial Neural Network (ANN), Extreme Gradient Boosting (XGB), and Multilayer Perceptron (MLP), to analyse gamma ray, resistivity, density, neutron porosity, acoustic, and photoelectric factor logs. The Concentration-Number (C-N) fractal model approach and log–log plots were also used to define geothermal features. In a study on machine learning models for classifying different rock types in the Sawan field of the Southern Indus Basin, it was discovered that sand (fine, medium and coarse) facies were most accurately classified (87–94%), followed by shale (70–85%) and siltstone facies (65–79%). The accuracy of the machine learning models was assessed using various statistical metrics, such as precision, recall, F1 score, and ROC curve. The study found that all five machine learning methods successfully predicted different litho-facies in the Lower Indus Basin. In particular, sand facies were most accurately classified, followed by shale and siltstone facies. The multilayer perceptron method performed the best overall. This multidisciplinary approach has the potential to save time and costs associated with traditional core analysis methods and enhance the efficiency of hydrocarbon exploration and development.