Supervised Classification Algorithms Research Articles

Efficient Detection and Localization of DoS Attacks in Heterogeneous Vehicular Networks

Vehicular communication has emerged as a powerful tool for providing a safe and comfortable driving experience for users. Long Term Evolution (LTE) supports and enhances the quality of vehicular communication due to its properties such as high data rate, spatial reuse, and low delay. However, high mobility of vehicles introduces a wide variety of security threats, including Denial-of-Service (DoS) attacks. In this paper, we propose effective solutions for real-time detection and localization of DoS attacks in an LTE-based vehicular network with mobile network components (e.g., vehicles, femto access points, etc.). We consider malicious data transmission by vehicles in two ways- using real identification (unintentional) and using fake identification (intentional). This paper makes three important contributions. First, we propose an efficient attack detection technique based on data packet counter and average Packet Delivery Ratio (PDR). Next, we present an improved attack detection framework using machine learning algorithms. We use some ML-based supervised classification algorithms to make detection more robust and consistent. Finally, we propose Data Packet Counter (DPC)-based, triangulation-based and measurement report based localization for both intentional and unintentional DoS attacks. We analyze the average packet delay incurred by vehicles by modelling the system as an <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"><tex-math notation="LaTeX">$M/M/m$</tex-math></inline-formula> queue. Our experimental evaluation demonstrates that our proposed technique significantly outperforms state-of-the-art techniques.

Improving the performance of the model developed from the classification of adulterated honey with different botanical origins based on near-Infrared hyperspectral imaging system and supervised classification algorithms

Honey is a high-value product that proscribes adulteration. Authentication of pure honey is an imperative act that can disregard the adverse impacts brought by fraudsters on honey producers and consumers. A reliable method that can detect food adulteration that is nondestructive, rapid, and efficient can help maintain product authenticity. This study investigated the potential of a near-infrared hyperspectral imaging (NIR-HSI) system combined with supervised classification techniques for the classification of honey samples from different botanical origins that were adulterated with sugar cane syrup at 20 % and 40 % w/w. Reflectance and transflectance modes of measurement were used to obtain honey spectral images from 864.53 nm to 1700 nm. Supervised classification algorithms, namely, stepwise discriminant analysis (SDA), support vector machine (SVM), and k-nearest neighbors (kNN) were used for the classification based on whole spectral data and feature selected spectral data. The best classification model from reflectance mode was observed from SDA based on a leave-one-out cross-validation (LOOCV) method with a test accuracy of 95 %. For transflectance mode, the classification model developed from SDA based on a LOOCV method provided better test set accuracy of 100 %. The classification performance levels of the models developed from SVM and kNN were lower with the use of the whole wavelength region. Moreover, the classification performance levels of SVM and kNN using the selected features by the SDA greatly increased, however, they were not as good as from using SDA alone. The classifier maps created using SDA and predicting the averaged spectra of all pixels in the image with 10-discriminant functions for each mode of measurement displayed a two-dimensional image that had a distinct colour pertaining to a particular class that could easily discriminate one class from another.

Generalized Zero-Shot Activity Recognition with Embedding-Based Method

Sensor-based human activity recognition aims to recognize the activities performed by people with the sensor readings. Most of existing works in this area rely on supervised classification algorithms, and can only recognize activities covered by the training data. Whereas, in many practical applications, while performing activity recognition, not only the activities covered by the training data, but also some previously unseen activities need to be recognized. In this paper, we study the problem of generalized zero-shot activity recognition. In this problem, the activities that need to be recognized contain both the activities covered by the training data and the previously unseen activities. We firstly give a formulation of this problem, and then propose an embedding-based method to address it. In this method, an embedding-compatibility model is learned. When performing activity recognition, the learned model and the calibrated stacking mechanism are employed. Extensive experiments on publicly available datasets demonstrate the effectiveness of our method.

MicroRNA mediated gene regulatory circuits leads to machine learning based preliminary detection of acute myeloid leukemia

Acute Myeloid Leukemia (AML) can be detected based on morphology, cytochemistry, immunological markers, and cytogenetics. MicroRNAs (miRNAs) influence key biological pathways in multiple haematological malignancies including AML. In this work, we have analysed the miRNome and the transcriptome of normal and AML samples and have identified the significant set of miRNA-target mRNA pairs present within AML- Peripheral Blood and AML- Bone Marrow samples from both tissue and cell lines. The miRNA target genes are further filtered based on their functional significance in AML system. These filtered genes constitute the set of selected miRNA target features, which have been finally used for developing machine learning based prediction tool, ‘TbAMLPred’ for preliminary detection of AML.This model implements both unsupervised clustering and supervised classification algorithms that would increase the reliability of prediction. Our results show that the selected miRNA target-based features can separate the control and disease samples linearly. Overall, we put forward ‘TbAMLPred’ for a non-invasive mode of preliminary AML diagnosis in future.Github link for accessing TbAMLPred: https://github.com/zglabDIB/TbAMLPred

Integrating Satellite Imagery and Ground-Based Measurements with a Machine Learning Model for Monitoring Lake Dynamics over a Semi-Arid Region

The long-term variability of lacustrine dynamics is influenced by hydro-climatological factors that affect the depth and spatial extent of water bodies. The primary objective of this study is to delineate lake area extent, utilizing a machine learning approach, and to examine the impact of these hydro-climatological factors on lake dynamics. In situ and remote sensing observations were employed to identify the predominant explanatory pathways for assessing the fluctuations in lake area. The Great Salt Lake (GSL) and Lake Chad (LC) were chosen as study sites due to their semi-arid regional settings, enabling the testing of the proposed approach. The random forest (RF) supervised classification algorithm was applied to estimate the lake area extent using Landsat imagery that was acquired between 1999 and 2021. The long-term lake dynamics were evaluated using remotely sensed evapotranspiration data that were derived from MODIS, precipitation data that were sourced from CHIRPS, and in situ water level measurements. The findings revealed a marked decline in the GSL area extent, exceeding 50% between 1999 and 2021, whereas LC exhibited greater fluctuations with a comparatively lower decrease in its area extent, which was approximately 30% during the same period. The framework that is presented in this study demonstrates the reliability of remote sensing data and machine learning methodologies for monitoring lacustrine dynamics. Furthermore, it provides valuable insights for decision makers and water resource managers in assessing the temporal variability of lake dynamics.

Semi-supervised networks integrated with autoencoder and pseudo-labels propagation for structural condition assessment

Insufficient labeled data of vibration measurement from unknown structural states cast great obstacle to attempt of extending the superiority of deep learning techniques into their rapid state evaluation. This study proposes novel semi-supervised networks for condition assessment integrated with deep autoencoder and pseudo-labels propagation. The architecture and mechanism of the workflow are elaborated. With sophisticated network design and novel strategies for improving performance, the proposed method succeeds to establish a systematic network embedded with simultaneously self-supervised autoencoder, optimized unsupervised fuzzy clustering and supervised classification algorithms in semi-supervised paradigm. Both numerical simulations and full-scale laboratory shaking table tests of a two-story building structure were conducted to validate its capability of identifying the post-earthquake damage scenarios. Driven by acceleration signals measured by merely single sensor, the prediction accuracy of proposed method achieved to be 0.95 in numerical validation and averagely 0.86 in laboratory case studies, respectively, confirming a powerful effectiveness and robustness.

Urban growth, changing relationship between biophysical factors and surface thermal characteristics: A geospatial analysis of Agra city, India

The kind of urbanization occurred in the last few decades has caused a massive reduction of urban green space raising the land surface temperature, drastic land use and land cover changes, and hampering environmental sustainability in Agra city. Examining urban growth trends and its significant impacts on surface and vegetation cover of Agra helps in assessing nature of spatial development process. The present study examines urban growth, bio-physical factors and surface thermal characteristics in the city. Study uses multiple geospatial techniques to analyse changes. To obtain spatio-temporal landscape dynamics multi-temporal Landsat TM/OLI satellite images were used for years 1991, 2001, 2011 and 2021 to classify LULC classes using maximum likelihood supervised classification algorithm. Based on the spatial drivers and LULC of 2001 and 2011; multi perception based artificial neural network model is applied to predict the urban areas in 2031. Study reports that the greatest loss was found for vegetation cover (VC); a total of 17.89 sq. km. (13.06%) green space lost during the study periods (1991–2021). A remarkable increased in urban area was noticed at a rate of 11.20% during 1991–2001 and 7.97% during 2001–2011. The study has much to offer to urban and environmental planning and sustainable urbanization.

Applications of Geospatial Techniques to Identify Landscape Changes and Urban Expansion of Dera Ghazi Khan City, South Punjab-Pakistan

A temporal analysis of landscape change patterns and consequent evaluation of urban expansion has been studied for the Dera Ghazi Khan city which is developing fast since 2000 as an emerging metropolis of South Punjab. The geospatial techniques particularly remote sensing augmented with GIS are vital tools and are mostly used in studying landscape changes and decision making for sustainable urban development. Four Landsat images (1991, 2001, 2011 and 2021) obtained from the free web of GLFC have been analysed. A supervised classification algorithm was applied to all images and results have been presented in both tabular and graphic forms. Present study, focuses upon two land-use classes i.e. built up and open areas, which will be monitored by land-use patterns of Dera Ghazi Khan city during 1991 to 2021 using satellite images. The graphical depiction of all the landscape changes, which have taken place in D. G. Khan city in the form of tables and maps provides detailed information about the degree and ratio of urban expansion and the changes in land use that occurred during the stipulated period.The classification was based on four categories i.e., urban land, bare soil, vegetation, and water; while maps were prepared to present the temporal changes in D. G. Khan city. A total of 52 sq./kms areas was selected for this study designated as the Dera Ghazi Khan city. According to the results, 33% of the urban area has been expanded in Dera Ghazi Khan from 1991 to 2021 and 19% vegetation cover has been reduced. It is a very alarming situation that agricultural land is decreasing because of urban expansion. The results revealed the significance of remote sensing and GIS in evaluating landscape changes.

Monitoring and forecasting of land use/land cover (LULC) in Al-Hassa Oasis, Saudi Arabia based on the integration of the Cellular Automata (CA) and the Cellular Automata-Markov Model (CA-Markov)

ABSTRACT This study sought to integrate the Cellular Automata-Markov Model (CA-Markov) and the Cellular Automata (CA) using sensing data for land cover maps for the years: 1988, 2000, 2013 and 2020 to monitor, detect, and predict the spatial and temporal of Land Use/Land Cover (LULC) change in Al-Hassa Oasis, Saudi Arabia. The maximum likelihood classifier (MLC) method was used as a supervised classification algorithm to super control classification coefficients. Training areas were extracted to classify green and water from Landsat images using The Natural Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI) indicators, respectively. The Markov form has been used Cellular Automata (CA)To create a Transition probability Matrix (TPM)), Transition Area Matrix (TAM) and Transition Suitability Maps (TSM), which is the main factor for obtaining to simulate the prediction of the future of land use/land cover in the years 2030 and 2050, while was used Cellular Automata-Markov Model (CA-Markov) To validate and Forecasting LULC. The results showed an expansion in urban areas by 63.5% equivalent to about 105.7 km2, and an increase in agricultural land by 6.6% equivalent to 10.7 km2 despite the loss of about 3.2% of the old agricultural land during the period 2000–2013 AD. Also, the area of marshes and water increased by about 41.7% equivalent to about 8.7 km, which led to the loss of 5.9% or the equivalent of about 139 km2 of arid lands during the past three decades. The projected LULC maps for 2030 and 2050 indicate that these patterns of arid lands that change into built-up areas and agricultural lands will continue over the next thirty years due to urban growth and the development of agricultural activities. The main strength of this study is the use of a relatively long period that allowed us to conduct a very good assessment of LULC in the study area, as well as to predict the different patterns for 2030 and 2050. Moreover, the results obtained in this study provide information to help decision makers to carry out effective practices for future planning and management of urban land growth and land use planning, especially in the Saudi Vision 2030.

Assessment of Aquifer Recharge Potential Using Remote Sensing, GIS and the Analytical Hierarchy Process (AHP) Combined with Hydrochemical and Isotope Data (Tamassari Basin, Burkina Faso)

In the Tamassari basin, the agricultural population is highly dependent on groundwater resources for its socioeconomic development. However, the decrease in rainfall in the region since the late 1960s and the demographic pressure on the land are significantly affecting groundwater recharge. In order to exploit this groundwater sustainably, it is necessary to identify potential recharge areas for a better capitalisation of this resource. The objective of this study is to map the recharge potential of the existing aquifers making use of remote sensing and GIS techniques and to make a validation based on chloride and tritium contents in the borehole water. The processing carried out on the Landsat 5 and Landsat 8 images combined with a digital elevation model (ALOS PALSAR), highlight the lithological, linear and topographical characteristics of the study area. In addition, various supervised classification algorithms were used to produce the most accurate land use map. Field campaigns were conducted to validate the thematic maps resulting from the geospatial data processing and to collect water samples for hydrochemical (chloride) and isotopic analysis (tritium). The analytical hierarchy process (AHP) was used to derive recharge factors weights. The resulting recharge map shows a perfect agreement between the recharge classes derived from spatial modelling and the tritium isotope analyses. This was not the case with the chloride contents, which showed a dispersion over all the recharge areas.

RVGAN-TL: A generative adversarial networks and transfer learning-based hybrid approach for imbalanced data classification

Imbalanced data distribution is the main reason for the performance degradation of most supervised classification algorithms. When dealing with imbalanced learning problems, the prediction of traditional classifiers tends to favor the majority class and ignore the minority class which is often much more important. Therefore, it is necessary to balance majority data and minority data before classification. A popular strategy for balancing the two data classes is synthesizing minority data. In recent years, generative adversarial networks (GAN) have shown great potential in fitting sample distributions. Based on this, this paper proposes a model combining improved GAN and transfer learning, RVGAN-TL, to solve the imbalanced learning problem of tabular data. As for the improvement of GAN, variational autoencoder (VAE) is used to generate latent variables with a posterior distribution as the input of GAN, and similarity measure loss is introduced into the generator to improve the quality of the minority data generated by GAN. In addition, a roulette wheel selection method is applied to the training data selection in GAN to rebalance data in the overlapping area. When data is balanced, the generated data is used as the source domain and the original data as the target domain, and the transfer learning method is used to train the final classifier. Experiments on 20 real datasets show that the classification performance of the proposed method is significantly improved compared with other popular methods.

Species-Specific Contribution to Atmospheric Carbon and Pollutant Removal: Case Studies in Two Italian Municipalities

In order to maximize ecosystem services (ES), a proper planning of urban green areas is needed. In this study, the urban greenery of two Italian cities (Milan and Bologna) exposed to high levels of atmospheric pollutants was examined. Vegetation maps were developed through a supervised classification algorithm, trained over remote sensing images, integrated by local trees inventory, and used as input for the AIRTREE multi-layer canopy model. In both cities, a large presence of deciduous broadleaves was found, which showed a higher capacity to sequestrate CO2 (3,953,280 g m2 y−1), O3 (5677.76 g m2 y−1), and NO2 (2358.30 g m2 y−1) when compared to evergreen needle leaves that, on the other hand, showed higher performances in particulate matter removal (14,711.29 g m2 y−1 and 1964.91 g m2 y−1 for PM10 and PM2,5, respectively). We identified tree species with the highest carbon uptake capacity with values up to 1025.47 g CO2 m2 y−1 for Celtis australis, Platanus x acerifolia, Ulmus pumila, and Quercus rubra. In light of forthcoming and unprecedent policy measures to plant millions of trees in the urban areas, our study highlights the importance of developing an integrated approach that combines modelling and satellite data to link air quality and the functionality of green plants as key elements in improving the delivery of ES in cities.

Diagnosis of Osteoarthritis at an Early Stage via Infrared Spectroscopy Combined Chemometrics in Human Serum: A Pilot Study

Methods applied for early diagnosis of osteoarthritis (OA) are limited. Early prevention and treatment can effectively reduce the pain of OA patients and save costs. The present study aimed to develop a rapid non-destructive detection method for early diagnosis of OA by evaluating infrared (IR) spectroscopy combined chemometrics. Our cohort consisted of (a) 15 patients with osteoarthritis (OA) and (b) 10 without clinical signs of the disease and they were used as controls. Attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy was used to investigate serum samples (50 µL) collected from these patients. A supervised classification algorithm namely discriminant analysis (DA) was applied to evaluate the diagnostic accuracy spectral processing and chemometrics analysis allowed for detecting spectral biomarkers that discriminated the two cohorts. About 250 infrared spectra were statistically important for separating the groups. Peaks at 1000 cm−1 in OA serum were associated mainly with C–O stretching vibration associated with the changes in the proteoglycan contents previously reported in OA. A good overall classification accuracy of 74.47% was obtained from the DA model. Our findings indicated that this discriminating model, which incorporated the ATR-FTIR spectrum, could provide a rapid and cost-effective blood test, thus facilitating the early diagnosis of human OA.

SHORELINE EXTRACTION USING TIME SERIES OF SENTINEL-2 SATELLITE IMAGES BY GOOGLE EARTH ENGINE PLATFORM

Abstract. In recent decades, global warming and sea level rise, population growth, and intensification of human activities, have directly affected the coasts and as such, their monitoring for the accretion and retreat are among the issues that are considered by the coastal countries. This study, compares two supervised classification algorithms for classifying Sentinel-2 satellite imagery for shoreline extraction. Median monthly images from 2020/01 to 2021/12 are taken and classified by Random Forest (RF) and Support Vector Machine (SVM) algorithms. By validating the maps, it is found that the RF algorithm has better accuracy and as such by averaging the accuracy of all maps, the overall accuracy (OA) values of 97.18% and the kappa coefficient (KC) of 0.97, and the mean overall accuracy and kappa coefficient of maps from SVM algorithm of 85.15% and 0.79, respectively, is obtained. After extracting the shorelines, the Digital Shoreline Analysis System (DSAS) is used to calculate the displacement rate. By calculating the Linear Regression Rate (LRR) factor, it is found that in 91% of transects (166 transects) we see the shoreline retreat to land. In 54% of them, the average rate of the retreat is 5.42 meters per year and in only 9% (16 transects) we see the accretion towards the sea.

Identifying Emeishan basalt by supervised learning with Landsat-5 and ASTER data

Multispectral-sensor images are advantageous in terms of discriminating major lithologies due to their high spatial resolution and intermediate spectral resolution, in addition to their low cost and high accessibility in comparison to hyperspectral images. In this study, Landsat-5 Thematic Mapper ™ and the Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) data—which are the most widely used multispectral data for the discrimination of the mixed rock units—are utilized to identify basalts in our study area. Further, prior knowledge regarding basalt-distribution areas in our study region is obtained from the geological-survey results conducted by the Sichuan Geological Survey at 2005, which is used as the reference of correction to assess our identified results. Small portions of this prior area of basalt distribution were verified through field checks, which were then determined as sites for use as training data for remote-sensing imagery. Three supervised-classification algorithms within ENVI 5.3—k-nearest neighbors (KNN), maximum likelihood classification (MLC), and support vertical machine (SVM)—were utilized for model identification. As a result, six models were constructed, including the KNN prediction of basalts by ASTER images, SVM prediction by ASTER, MLC prediction by ASTER, KNN prediction by Landsat-5 images, SVM prediction by Landsat-5, and MLC prediction by Landsat-5. The performances of the six models, in terms of precision and accuracy, show that the optimum model is Landsat-5 by SVM, with a precision of 70.92% and accuracy of 99.97%, followed by the ASTER by SVM model, with a precision of 67.72% and accuracy of 99.89% and the Landsat-5 by KNN model, with a precision of 57.23% and accuracy of 99.85%.

SPATIO TEMPORAL MAPPING AND MONITORING OF LAND COVER DYNAMICS OF ISLAMABAD USING MULTI-TEMPORAL REMOTE SENSING DATA

Land cover resources are changing continuously both spatially and temporallywhereas the accurate and timely information regarding their dynamics is very important. Spatialgrowth is an expected phenomenon particularly in fast growing metropolitan areas like Islamabad, thecapital of Pakistan. In this study spatial and temporal dynamics of land cover changes were quantifiedfrom 1975 to 2010 using Landsat images, a supervised classification algorithm and the changedetection technique with Geographical Information Systems were used. The overall accuracy of theland cover derived from satellite images was from 80 to 95%. The land cover maps showed thatbetween 1975 and 2010 the amount of urban or developed land increased from 50 km2to 332 km2 of the total area, while vegetation, including Margalla Hills National Park decreased from 465 km2to 145 km2. The results quantified the land cover change pattern in the city and showed a potential of multitemporal Landsat data to provide an accurate, cost-effective means to map and analyze changes in land cover over time

A Comparison of the Various Methods for Selecting Features for Single-Cell RNA Sequencing Data in Alzheimer's Disease.

The high-throughput sequencing method known as RNA-Seq records the whole transcriptome of individual cells. Single-cell RNA sequencing, also known as scRNA-Seq, is widely utilized in the field of biomedical research and has resulted in the generation of huge quantities and types of data. The noise and artifacts that are present in the raw data require extensive cleaning before they can be used. When applied to applications for machine learning or pattern recognition, feature selection methods offer a method to reduce the amount of time spent on calculation while simultaneously improving predictions and offering a better knowledge of the data. The process of discovering biomarkers is analogous to feature selection methods used in machine learning and is especially helpful for applications in the medical field. An attempt is made by a feature selection algorithm to cut down on the total number of features by eliminating those that are unnecessary or redundant while retaining those that are the most helpful.We apply FS algorithms designed for scRNA-Seq to Alzheimer's disease, which is the most prevalent neurodegenerative disease in the western world and causes cognitive and behavioral impairment. AD is clinically and pathologically varied, and genetic studies imply a diversity of biological mechanisms and pathways. Over 20 new Alzheimer's disease susceptibility loci have been discovered through linkage, genome-wide association, and next-generation sequencing (Tosto G, Reitz C, Mol Cell Probes 30:397-403, 2016). In this study, we focus on the performance of three different approaches to marker gene selection methods and compare them using the support vector machine (SVM), k-nearest neighbors' algorithm (k-NN), and linear discriminant analysis (LDA), which are mainly supervised classification algorithms.

Analysis of Learning Behavior Characteristics and Prediction of Learning Effect for Improving College Students’ Information Literacy Based on Machine Learning

Information literacy is a basic ability for college students to adapt to social needs at present, and it is also a necessary quality for self-learning and lifelong learning. It is an effective way to reveal the information literacy teaching mechanism to use the rich and diverse information literacy learning behavior characteristics to carry out the learning effect prediction analysis. This paper analyzes the characteristics of college students’ learning behaviors and explores the predictive learning effect by constructing a predictive model of learning effect based on information literacy learning behavior characteristics. The experiment used 320 college students’ information literacy learning data from Chinese university. Pearson algorithm is used to analyze the learning behavior characteristics of college students’ information literacy, revealing that there is a significant correlation between the characteristics of information thinking and learning effect. The supervised classification algorithms such as Decision Tree, KNN, Naive Bayes, Neural Net and Random Forest are used to classify and predict the learning effect of college students’ information literacy. It is determined that the Random Forest prediction model has the best performance in the classification prediction of learning effect. The value of Accuracy is 92.50%, Precision is 84.56%, Recall is 94.81%, F1-Score is 89.39%, and Kapaa coefficient is 0.859. This paper puts forward differentiated intervention suggestions and management decision-making reference in the information literacy teaching process of college students, with a view to adjusting the information literacy teaching behavior, improving the information literacy teaching quality, optimizing educational decision-making, and promoting the sustainable development of high-quality and innovative talents in the information society.Our work involving research of the thinking and direction of the sustainable development of information literacy training proved to be encouraging.

On the benefits of machine learning classification in cashback fraud detection

Technology development has been getting more advanced and greatly facilitated human life. One of them is machine learning automation which has been proven to be consistent for doing various computations against extensive data such as transaction data in the e-commerce area. Seeing this opportunity, we implemented the machine learning approach to detect fraudulent cashback transactions in e-commerce that are currently rife in Indonesia. The training data used to build the machine learning model were the transaction data from one of the leading e-commerce in Indonesia that had been processed. The supervised classification algorithms used were K-Nearest Neighbor (k-NN), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM). In the end, the best steps and methods that could be taken against fraudulent cashback activities in the future are shown in this paper.

Assessment of land use and cover in the Sucuru Watershed using Google Earth Engine

Human activities modify the natural characteristics of numerous watersheds worldwide. Google Earth Engine provides tools for the analysis of land use and natural resources. In this work, we classify current land use and cover in the Sucuru watershed, Paraíba, Brazil. We compared the accuracy of five supervised classification algorithms of Google Earth Engine. Classifiers based on Decision Trees, such as the Classification and Regression Trees (CART) and Random Forest (RF), showed the best accuracy and visual inspection values. The Google Earth Engine is a powerful tool for analysis of large-scale environmental data, monitoring land use changes, and providing information for sustainable management.