Search

Online social media platforms such as YouTube have grown in popularity in recent years, providing as critical routes for content creators, advertisers, and platform administrators to access a large worldwide audience. Accurate video popularity prediction is critical for improving content strategy, increasing user engagement, and increasing video visibility. This study intends to improve video popularity prediction accuracy by presenting a fresh methodology that goes beyond traditional analysis. To construct a more effective prediction model, the suggested method combines examining metadata, such as title, description, and channel title, with thumbnail analysis. The project aims to dramatically increase prediction accuracy by extracting relevant information from both metadata and thumbnails. To do this, multiple classification techniques, including Support Vector Machine, Gradient Boosting, Random Forest, Extreme Gradient Boosting, and K-Nearest Neighbors, were used to train and forecast the model. Notably, when it came to feature extraction, the Gradient Boosting and Random Forest algorithms achieved amazing accuracy of 97% when using all extracted and original data, and 82% and 92% when using only extracted features. In comparison, previous studies employing simply extracted features obtained only 74% accuracy. These findings demonstrate the potential benefit of integrating metadata and thumbnail data research to get insights into the elements that drive video popularity and lead to more accurate predictions. [ABSTRACT FROM AUTHOR]

Mathematical programming techniques determine the optimal functional configuration of a Wireless Sensor Network (WSN). However, the high computational complexity of these techniques has been regarded as a Non-Polynomial problem. Hence, Machine Learning (ML) approaches are used for WSN parameter prediction with improved accuracy and less computing complexity. This study aims to predict the reliability of clustering algorithms in WSNs using supervised ML classification approaches. ML algorithms, namely, decision tree and K-nearest neighbor classifications are implemented on the obtained dataset to predict the reliability of clustering in WSNs on the basis of high packet delivery ratio, low delay, and high residual energy. The reliability parameter accepts binary values, where zero represents the absence of reliability and one represents the presence of reliability metrics. Then the predictions from each classifier are compared with one another. The results show that the accuracy rate of the decision tree is 97.98% and the error rate is 2.02% which shows that the decision tree performs better than the K-nearest neighbor in predicting the reliability of the proposed approach. [ABSTRACT FROM AUTHOR]

Aim-The aim of the study is to perform innovative detection of lung cancer using the Decision Tree classifier and also comparing it with the K-Nearest Neighbor classifier. Materials and methods-The lung cancer database includes records 32 samples (patients) and is therefore compiled with 32 ∗ 6 numerical data. The G power for samples is calculated from clincalc which contains two different groups where group 1 is taken as (n1=16) and for group 2 (n2=16), alpha (0.05), power (80 %) and enrollment ratio. Results-Implementation based on the MATLAB toolbox is done on CT lung images and the classification of these images is performed. In the proposed method, D-Tree numbers of samples are selected for each iteration and 90 % phase accuracy is obtained. There is a significant difference in Accuracy rate (P=0.072). Conclusion-The results of the study suggest that the accuracy of the decision tree algorithm is significantly higher than the nearest k-neighboring algorithm. [ABSTRACT FROM AUTHOR]

The research aims to enhance liver tumour prediction using an unique K-Nearest Neighbor (KNN) machine learning technique. Materials and Methods: The research employed 52 samples with two groups of algorithms with a g-power of 80% and datasets from different online sources with current study results, threshold 0.05 percent, confidence interval 95 percent mean and standard deviation. Logistic regression predicts liver cancers with 71.11 percent accuracy, thus this work needs to improve innovative liver tumour prediction using KNN. Results: The innovative K-Nearest Neighbor method achieved 74.29 percent accuracy in liver tumour prediction, with a significant value of 0.045 (p<0.05) and a 95% confidence range. This research found that KNN predicts liver tumours better than Logistic regression. [ABSTRACT FROM AUTHOR]

The primary goal of this study is to perform leaf disease diagnosis using K-Nearest Neighbors classification (KNN) and compare its performance against that of Support Vector Machine (SVM) technique to increase accuracy. Materials and Techniques SVM was used in this study to detect leaf illness by sampling 10 samples (N=10), and K-Nearest Neighbors was used to classify the samples (N=10). With a sample size of 20 for each group and a pretest power of 80%, two groups are analyzed statistically. Both strategies' accuracy is evaluated. The Support Vector Machine algorithm outperforms the KNN approach in terms of accuracy (96.25% vs. 83.3%), and independent samples T-tests show a statistically significant difference between the two algorithms' accuracy at p=0.37 (p0.05). Conclusion: According to the results, the Support Vector Machine technique for identifying leaf diseases appears to be noticeably superior to K-Nearest Neighbors categorization. [ABSTRACT FROM AUTHOR]

The purpose of this study effort is to undertake the comparative analysis of detection rate in real time innovative face recognition systems using K-Nearest Neighbor algorithm in comparison with Linear Discriminant Analysis method. Materials and Methods: The sample size of a real time face recognition system with increased detection rate utilising Linear Discriminant Analysis algorithm in comparison with K-Nearest Neighbor algorithm was 20 by considering 10 numbers of samples for each approach with pre-test power of 80 percent. Findings: The detection rate of K-Nearest Neighbor method is 93.0258 percent while results of Linear Discriminant Analysis technique detection rate are 89.762 percent. There is a significant variation in Detection rate (P=0.0421). The findings obtained were judged to be error-free as it was having the significant value p<0.05 in SPSS statistical analysis. Conclusion: K-Nearest Neighbor method depicts the greater performance in discovering the detection rate in a real-time face recognition system when compared to Linear Discriminant Analysis algorithm. [ABSTRACT FROM AUTHOR]

The K-Nearest Neighbor approach outperforms the Fuzzy Algorithm in terms of accuracy during quick data retrieval for an innovative car parking system. The Components and Techniques: K-Nearest Neighbor Algorithm enables 20-sample database vehicle parking system, which improves innovative vehicle parking system's data retrieval accuracy and speed. The implementation was accomplished by developing a Python web application with the aid of the anaconda navigator. As a result, a car parking database is constructed using the K-Nearest Neighbor Algorithm, which improves efficiency by 87 percent in terms of secure accessibility compared to the FuzzyApproach, which improves efficiency by only 75 percent. The length of time it takes to log in and access the database has been analysed using SPSS (Statistical Package for the Social Sciences) with the independent variables of time and size and a significance level of p 0.05. K-Nearest Neighbor Algorithm, which is more significant than a Fuzzy Approach in terms of Accurate efficiency, has been employed in a protected system to improve the car parking database. When data is a dependent variable and sample size is an independent variable, SPSS analysis may show how reliable the data is. [ABSTRACT FROM AUTHOR]

The goal of the research is detection of disease in fresh fruits. The sensitivity and accuracy was calculated and compared with the k-nearest neighbor (KNN) classifiers. Materials and Methods: The research's data set was obtained from the Kaggle open access dataset. The research includes two groups, in which CNN is group 1 and KNN is group 2. Kaggle database is used for collecting different fruit images. The sample size was calculated using G power (80%) and alpha (0.05) and it is found to be 20 samples for each group. Result : CNN has achieved 98 % of accuracy and 1.0477 standard deviation while KNN method has achieved 90 % of accuracy and 0.65071 standard deviation. The Novel CNN has significant accuracy of p<0.05 (2 tailed). Conclusion: It is concluded that the detection of fruit disease using Convolution Neural Network appears to be significantly better compared to k-Nearest Neighbour Classifiers. [ABSTRACT FROM AUTHOR]

The main goal of the study was to detect and classify spot diseases in fruits using artificial neural networks (ANN) more accurately than K-nearest neighbors. The dataset available on Github was used to collect samples of 40 different RGB photos. Using Clinicalc.com, sample size was determined with a pretest power of 80%. The accuracy of these samples was evaluated using ANN and ANN as experimental and control groups, respectively. Based on the accuracy, the effectiveness of ANN and KNN for fruit RGB photos is examined. For artificial neural network and ANN, the accuracy rates are 92.08% and 87.14%, respectively. According to statistical analysis, a significant value of p 0.05 for ANN and independent sample t-test can bring more accurate and desirable results for RGB images. This study finds that ANNs outperform ANN algorithms. [ABSTRACT FROM AUTHOR]

The purpose is to use attenuation measurements of dynamic tuning parameters to classify and predict floods in advance using rainfall data patterns in India. This may be done by using rainfall data patterns in India. We use two distinct categorization approaches in order to achieve the best possible outcomes. Each of these approaches has a sample size of five, a G power of eighty percent, a threshold of five percent, a confidence interval (CI) of ninety-five percent, as well as a mean and standard deviation. Using information on previous rainfall, we conducted this investigation in which we examined the accuracy of forecasts generated by two distinct methods: Random Forest and K-Nearest Neighbor. The findings demonstrated that Random Forest had a performance that was 51.47 percentage points higher than that of the K-Nearest Neighbor algorithm (50.35 percent). It has been demonstrated that Rainfall-based Random Forests are superior to K-Nearest Neighbor algorithms when it comes to flood prediction. [ABSTRACT FROM AUTHOR]

The primary goal of this research is to use J48 and K-Nearest Neighbor (KNN) classifiers to predict arrhythmia using the MIT-BIH database. With an alpha of 0.05, 95% confidence interval (CI), 80% power, and an enrollment ratio of 1, the proposed study used the J48 and KNN machine learning algorithms to predict arrhythmia using data from the MIT-BIH dataset consisting of healthy (n=65) and arrhythmia (n=65) ECG signals collected from IEEE dataport in.XLSX format. WEKA 3.8.5, a data mining tool, was used to distinguish between those with arrhythmia and those without. IBM SPSS version 21 was used for the statistical analysis. There was no discernible difference (p=0.025) between the J48 and KNN classifiers. Using WEKA's 10-fold cross validation to train, test, and verify the classifiers, we find that the J48 classifier is more accurate at classifying data (89.80 percent) than the KNN classifier (87.64 percent). [ABSTRACT FROM AUTHOR]

This paper's objective is to improve accuracy in plant leaf disease detection using Machine learning techniques such as K-Nearest Neighbors classifier (KNN) with a sample size of 10 and Novel Convolution Neural Network (CNN). This article consists of 2 groupsi.e Convolutional Neural Network (CNN) and K-Nearest Neighbors Classifier with a sample size of 10 for each group. G Power software is used to determine sample size with a pretest power value is 0.8 and alpha 0.05. The accuracy evaluated using both algorithms is compared and CNN appears to be the most significant with an accuracy of 92.48% and KNN with an accuracy of 74.14% with p=0.041. An Independent sample t-test has been used and analyzed with a confidence level of 95 percent. Results show that the CNN algorithm is significantly better than KNN in detecting plant leaf disease. [ABSTRACT FROM AUTHOR]

Using KNN Classifier against Linear Regression, we want to identify the flower and then publish its name beside it on the same picture. The 17flowers dataset at VGP (Visual Geometry Group) at the University of Oxford consists of a total of 1 360 pictures. Group 1 in Visual Geometry is classified using KNN, while Group 2 is classified using Linear Regression, both with a sample size of (N=5). This paper may be put into practise with the help of jupyter notebook. The effectiveness of the KNN Classifier is measured in terms of the accuracy values computed in comparison to the G power value, which is set at 80%. The KNN Classifier has a 72% success rate on average, whereas the Linear Regression model only has a 67% success rate. This research shows that when it comes to flower categorization, KNN Classifier (mean=72) is a clear winner over Linear regression (mean=67). [ABSTRACT FROM AUTHOR]

A comparative analysis of a novel convolutional neural network and k-nearest neighbors is performed to identify fingerprints in low-resolution photographs with high accuracy and sensitivity. Materials and methods: In this study, two groups were compared: a novel convolutional neural network (N=10) and a k-nearest neighbor algorithm (N = 10). Total sample size was estimated with an alpha of 0.05, an enrollment rate of 0.1, a confidence interval of 95%, and a power of 80% using G Power software. Results: The proposed method shows that the novel convolutional neural network achieves 93% accuracy and 90% sensitivity, while the K-Nearest Neighbor classifier achieves 89% accuracy and 87% sensitivity. The obtained accuracy rate was significant (p=0.005), and in the SPSS statistical analysis, the specificity value was (p=0.006). Conclusion: The new CNN classifier provides significantly better fingerprinting results than the K-Nearest Neighbors classifier. [ABSTRACT FROM AUTHOR]

An evaluation of SVM and a novel K-Nearest Neighbor classifier for fingerprint verification improvements. Methods and Materials In this article, we compare and contrast the Novel K-Nearest Neighbor Method with a Support Vector Machine. The recommended procedure is applied to a total of 15 samples, 15 from each of the two groups, and the G power is set to 80%. Results: The innovative K-Nearest Neighbor Method method predicts biometrics with an accuracy of 87 percent, which is much higher than the 82 percent accuracy of Support Vector [1]. Here, we calculated a pretest power of 85%, with a sample size of 15 in one group and 30 in total. The mean accuracy for Novel K-Nearest Neighbor (87.12%) and the SVM method (81.85%) are quite similar, however the standard deviation for Novel K-Nearest Neighbor (1.29517) is lower (2.45449) While the Novel K-Nearest Neighbor Method and SVM algorithms achieve similar mean accuracy (87.12 and 81.85 percent, respectively), the standard deviation for the Novel K-Nearest Neighbor Method is somewhat less (1.29517 vs. 2.45449), which is statistically significant at the 0.001 level (p 0.05). The Novel K-Nearest Neighbor Method Algorithm is discovered to be much superior than the SVM algorithm in this research for predicting biometric recognition. [ABSTRACT FROM AUTHOR]

This research endeavours to evaluate the efficacy of the support vector machine (SVM) method against that of the K-Nearest neighbour Network (KNN) approach in removing thick fog from images. We gathered the Realistic Single Image Dehazing (RESIDE) dataset from the kaggle.com database. For each group's sample size of 22, the suggested SVM method is compared against the current KNN techniques. There are 44 samples altogether. Our suggested process include measures to actually accomplish the elimination of noise in photos. Metrics of performance, such precision and sensitivity, are computed. The suggested support vector machine (svm) has an accuracy and sensitivity of 94.9 and 38.5%, respectively, whereas k-nearest neighbour had 93.7 and 37.5%. Using the G power instrument and setting the minimum power to 0.8, we can determine how many samples are needed for this study. When comparing the SVM and KNN algorithms, the significance level reached was (p 0.05). Support vector machines provide far superior performance over the standard k-nearest neighbour, as shown both by experimental findings and Independent sample T-test statistical analysis using SPSS software. [ABSTRACT FROM AUTHOR]

In this study, we compare the performance of two different algorithms for recognising irises: the K-Nearest Neighbor (KNN) strategy and the Support Vector Machine (SVM) approach. The Components and Procedures: Training and test sets were created from the whole research sample. In order to enhance the accuracy rate of the KNN and SVM algorithms, samples were collected from 100 people, with 5 eye scans from each individual (Group 1=250 and Group 2=250). KNN has a 93.34% accuracy rate, whereas SVM achieves just an 89.735% success rate. KNN has a 94.876 percent Precision rate, whereas SVM only has an 89.032 percent Precision rate. Recall for KNN is 94.298 percent, while recall for SVM is 90.123 percent. G-Power is assumed to be 0.8 due to the statistically significant discrepancy in accuracy rates (P=0.0427). The results of the iris recognition system show that the KNN algorithm is superior than the SVM technique. [ABSTRACT FROM AUTHOR]

The primary objective of this research is to diagnose lung cancer more accurately utilizing the Support Vector Machine (SVM) method and K-Nearest Neighbors classification (KNN) technique. Materials and Methods - SVM was used in this study to detect lung cancer using 10 samples (N=10), while KNN was used to classify the samples using 10 samples (N=10). The accuracy of the two approaches is contrasted. Results: The SVM method outperforms the KNN approach in accuracy (96% vs. 68%). Conclusion - It is clear from the results of the experimental testing that SVM performs better on lung imaging datasets than existing classifiers. The findings indicate that, for the identification of lung cancer, the SVM algorithm appears to be significantly better (p<0.05). [ABSTRACT FROM AUTHOR]

Recent technology breakthroughs have made it possible to automatically detect skin cancer using machine learning algorithms. Nonmelanoma skin cancers of the head and neck may now be diagnosed precisely and promptly thanks to the use of machine learning (NMSC). The researchers set out to find real-world medicinal uses for machine learning. K-Nearest Neighbor (KNN) and Logistic Regression were used as unique classification methods in this research with two distinct sample sizes (LR). Groups 1 and 2 of the KNN and LR algorithms had g-power values of 80%, and the squamous cell cancer images were collected from a wide range of internet resources using up-to-date research outcomes, a 0.05 percent threshold, a 95% confidence interval for the mean, and a standard deviation. Researchers often resort to methods that depend on machine learning because of how challenging it is to diagnose these illnesses. The KNN method achieved 86.11 percent accuracy whereas the LR algorithm only achieved 82 percent accuracy, with a significant value of 0.035 (p0.05) and a 95 percent confidence interval. Recommendations for POC skin cancer prevention and early detection are guided by a thorough analysis of the available data using the Machine Learning Approach. [ABSTRACT FROM AUTHOR]

To enhance lung cancer CT scan picture categorization utilising algorithms. Materials and Methods: Training dataset (n=185 [70%]) and test dataset (n=80 [30%]) are created from 265 lung cancer CT scan pictures. Weka tool uses Decision tree algorithm and KNN classifier for classification. Results and Discussions: Decision tree method classified lung cancer CT scan pictures with 99 percent accuracy and KNN classifier with 92 percent. Statistically, decision tree algorithms and KNN classifiers meet independent sample T-Test value (p<0.005) with G-power estimate of 0.8. Conclusion: The Decision tree approach outperforms the KNN classifier in lung cancer CT scan image categorization. [ABSTRACT FROM AUTHOR]

The objective of this research is to predict the loan sanction by comparing accuracy using Support Vector Machine (SVM) and K-Nearest Neighbour algorithm (KNN). Materials and Methods: Total sample size of this study is 40. Group 1 has 20 samples for KNN and group 2 has 20 samples for SVM. These groups were analyzed by an independent sample T-test with a pretest power of 80%. Results: KNN algorithm achieves accuracy of 70.90% and SVM achieves accuracy of 75.17%. The obtained significance value is 1% which is less than 5%. Conclusion: SVM achieves significantly better accuracy when compared with KNN algorithm. [ABSTRACT FROM AUTHOR]

The primary aim of the study was to assess the potential of two Algorithms (Novel K-Nearest Neighbor and Decision Tree) in classifying audio files according to genre, based on the features extracted from them. To this end, a dataset of 1000.wav files (GTZAN) was utilized for collecting and training the models with Mel-frequency cepstral coefficients(MFCCs), spectral centroid, zero-crossing, Chroma frequencies, and spectral roll-off via Python Librosa library. Sample sizes of 20 for both groups faced comparative analysis using SPSS software. Additionally, feature selection techniques identified characteristics that are effective for increasing accuracy. Results showed that KNN yielded 64.33%, outperforming the Decision Tree algorithm at 58.23%. Thus, it can be concluded that KNN is more proficient than a Decision tree for Music Genre Classification. [ABSTRACT FROM AUTHOR]

After lung cancer, prostate cancer is the most common form of cancer among men. World Cancer Research Fund International estimates that 1,414,259 people will be diagnosed with cancer in 2020, with 375,304 people losing their lives to the disease [1]Predicting the onset of Prostate Cancer at an early stage is, thus, essential. One of the major problems in Pca is diagnosing the disease as it does not show any early signs or symptoms. There are various diagnosing techniques for Pca. These techniques are very painful for the patients; one is DRE (Digital Rectal Examination). This is an ideal situation for using machine learning techniques to find an answer. Various supervised learning techniques (such as Random Forest, K-nearest Neighbor, SVM (Support Vector Machine), Logistic Regression, MLP (Multi-layer Perceptron), RNN (Recurrent Neural Network), and Naive Bayes) will be compared and contrasted in this study to make accurate predictions about prostate cancer. The primary goal of the research is to identify the most effective strategy for the prediction of prostate cancer for future use. Data for the study was used from the open-access online site of 100 patients. These Methods were evaluated basedon their precession, F1-Score, accuracy, recall and AUC. The results show that MLP, Logistic Regression, RNN performed well with 93% accuracy. Therefore, we may conclude that this method has huge benefits for the detection of prostate cancer. [ABSTRACT FROM AUTHOR]

Early Parkinson's disease (PD) diagnosis is essential for preventing the illness's development and providing patients with the opportunity to take disease-specific therapy. Speech signal processing is widely used and has gained attention recently. In this project, we carry out a comparative analysis for effective dysphonia (a speech disorder) and Parkinson's disease spotting using Machine Learning methods. To distinguish between PD patients and normal individuals. We used Naïve Bayes (NB), Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), Logistic Regression (LREG), and K-Nearest Neighbor (KNN) algorithms to demonstrate a robust detection process. According to experimental findings, the KNN outperformed other ML algorithms. The results of the trial conducted at UCI contain data from 31 patients, 23 of whom were diagnosed with PD. After implementing these algorithms, we found that our accuracy rates for these algorithms are NB (86.95%), RF (85.77%), SVM (79.18%), LREG (91.56%), DT (87.16%), and KNN (97.67%). [ABSTRACT FROM AUTHOR]

Indonesia is the largest corn exporter in the world. Corn (Zea mays I.) Problems in determining the selection of corn seed to replant, especially corn in East Nusa Tenggara, are still a critical issue. The things that affect the quality of corn are found: the seeds are damaged, the seeds are dull, the seeds are dirty, the beans are broken due to the drying process, and the shell of the corn. The determination of the quality of corn grains usually is done manually with visual observation. The manual system requires a long time and produces good quality products that are not consistent due to the limitation of visual fatigue and differences in the perception of each observer. This research uses image texture extraction comparison with statistical methods I orde (color moment) and orde statistics II (GLCM) to select the corn seed. Orde statistics I (color moment) shows the emergence of the value of the degree of gray probability pixels in an image, while orde statistics II (GLCM) shows the relationship between two probability pixels forming a concurrence matrix from the image data. This research is expected to help the process of classification in determining the corn seed. The algorithm k of the nearest neighbor (K-NN) who used to research the classification of the object of the image that will be examined. The results of this study successfully performed using k-Nearest neighbor (k-NN) with a distance of euclidean distance and k=1 with the extraction of the color moment got the highest accuracy is 88%, and the extraction GLCM to characterize the homogeneity of 75.5%, correlation of 78.67%, a contrast of 65.75% and energy of 63.82% with an average accuracy of 70.93%. [ABSTRACT FROM AUTHOR]

The color characteristics of pesticide vegetables can be seen with the naked eye, but there are inaccuracies in the assessment that continue to be carried out so that time efficiency is not met. KNN processes data from the color of vegetables with pesticides by classifying all images of vegetables. The results of the analysis showed that in this test the authors found that cauliflower vegetables did not contain pesticides, where the results obtained from color extraction were carried out from training results, and the test stated that cauliflower vegetables were vegetables that did not contain pesticides or were suitable for consumption. The results of this test use 30 test data with attributes and 3 species in the data classification. The test results show the K-Nearest Neighbor method in data classification has a good percentage of accuracy when using random data. The percentage of variation in the value of K K-Nearest Neighbor 3,5,7,8,9 as a percentage of 100%. [ABSTRACT FROM AUTHOR]

Nowadays there are a lot of cafes in Indonesia, especially in Jakarta, which is growing very rapidly. Youngsters and adults come just for hang out with their friends or families. Of course customers have favorite and non-favorite foods and drinks in the menu. To increase and maintain the customer numbers, a café must keep being creative and monitoring their menu. This research helps the Kopi Krintji Café in innovating new variants of food and drink to be added to their menus to comfort customers. This research utilizes the Naïve Bayes and K–Nearest Neighbor (KNN) algorithms to classify the Kopi Krintji Cafe's menu. The data for experimental purpose is collected from daily transactions for the period of January – May 2021. The raw data are then preprocessed and labeled as 'selling' and 'not selling' menu. The experimental results show that KNN algorithm provides an accuracy value of 98%, precision and recall, each of 98% with 10-fold cross validation test of 98% accuracy in predicting whether a menu is 'selling' or not. On the other hand, Naïve Bayes algorithm provides 97% accuracy, 95% precision and 93% recall while accuracy during the validation using k-fold cross validation with values of k are 10, 15, 20 and 25 is 96%. The Kopi Krintji Café management may consider this classification results as a basis when creating a successful new menu in the future. [ABSTRACT FROM AUTHOR]

A broad increase in data consumption in society and industry trigger network operators looking to upgrade their metro networks with higher bandwidth requirements. Service providers and operators are challenged to find a simple, the most efficient and cost-effective way of meeting the demand with new speeds and standards on the horizon. Distributed Denial of Service (DDoS) attack is a cyber-attack that uses a technique to flood the server, the system, or network of the targeted attack with unwanted traffic. The occurrence of DDoS attack on the metro networks can make the operating system unable to operate properly and even crash. DDoS can be prevented by monitoring traffic regularly, increasing server resource capacity and implementing multiple protection strategies. This paper investigates DDoS attacks by utilizing Information Gain feature Selection method based on metro network expert's opinion. The main aim is to improve the detection accuracy as such may help the metro network optimally provides the necessary bandwidth. Then, Naïve Bayes and K-Nearest Neighbor (KNN) classifiers are considered for evaluating the selected features as basis for the attack detection. Experimental results using CICIDS-2018 dataset show that KNN outperforms Naïve Bayes classifier with the accuracy level of 99%. [ABSTRACT FROM AUTHOR]

Attendance is an important factor in assessing employee discipline for a company. Discipline issues in recording employee attendance are an important concern for companies that have many branches or outlets because the company cannot directly monitor their employees. Some companies usually use finger print machines and some even use manual methods. This method is a problem for several companies that have several outlets or branches to record attendance data centrally and accurately. As a solution that is with the application of employee attendance that applies geofencing and face recognition as security in recording the attendance of employees who are at the company's outlets or branches. Geofencing is a technology for conducting remote surveillance of a predetermined area. While face recognition is a technology from a computer to recognize a person's face. In this study using the haversine formula, Euclidean distance, and KNN (K-Nearest Neighbor) algorithm in determining the last location and face recognition of employees by comparing face data that has been registered in the database. Users who record attendance outside the radius cannot record attendance. The face recognition accuracy rate has a percentage of 98%. [ABSTRACT FROM AUTHOR]

Classification analysis is a method used to predict classes or categories based on data class that has been determined previously. The k Nearest Neighbor (kNN) method is one of the most widely used methods in classification analysis due to the ease and simplicity of the algorithm. This strategy, however, is not without flaws; the determination of k, which is difficult to determine, and the determination based on the value of k results in data being classified into a particular class, even though it has a long distance. The Dependent Nearest Neighbor (dNN) method is a method that determines the nearest neighbor based on similarities and dependencies. In the dNN method, the closest selected neighbor is a sample in the Dependency Region (DR). DR is an area formed from the parameters of the radius and an angle. This study compares the performance generated by kNN and dNN using three datasets. Based on the analysis that has been done, the accuracy produced by the dNN method is greater than the kNN method. [ABSTRACT FROM AUTHOR]

The diverse motifs of Lampung Batik require an automated identification process using digital image processing technology. This research aims to identify Lampung Batik motifs by utilizing computer technology, particularly in digital image processing, to assist in recognizing these motifs. In this era of digitization, there are various object recognition methods, one of which is image recognition that utilizes GLCM feature extraction to obtain the characteristics of Lampung Batik motifs. After the feature extraction process on Lampung Batik images is completed, this data is used for classification. This study employs the K-Nearest Neighbor and Naïve Bayes classification methods, with a total dataset consisting of 300 images, comprising 150 images of "siger" motifs and 150 images of "sembagi" motifs. In testing with the K-Nearest Neighbor method, the highest accuracy is achieved at K = 3, approximately 77% for the training data and 81% for the testing data. Meanwhile, in testing with the Naïve Bayes method, an accuracy of around 96% is attained for both the training and testing data. [ABSTRACT FROM AUTHOR]

One of the things that is highly expected when collecting data is to produce complete data. In research, incomplete data will affect the results obtained. This is due to the non-maximum process carried out in the research. A dataset is a collection of data information that has been stored for a long time and becomes a large pile of data. Not infrequently, the dataset used in the research data presented is not complete. The missing value problem can be solved using data mining techniques. Data mining is the process of extracting information from a collection of data already stored in the data warehouse. Classification is the process of finding a common identity among different entities and classifying them into appropriate classes. Classification of large and complex data if performed manually would be difficult and take a long time. The K-Nearest Neighbor Imputation algorithm is a system that uses a supervised learning algorithm and aims to find new data patterns by connecting existing data patterns with new data. The conclusion that the authors drew is that the application of data mining to search for lost data using the K-Nearset Neighbor Imputation (KNNI) method is a process for generating new knowledge in the form of a comparison between the factors that influence the search for the missing data. The results of data mining using the K-Nearest Neighbor Imputation (KNNI) method are an arrangement of sequences of activities that support each other in the process. [ABSTRACT FROM AUTHOR]

Over the past few decades, there has been a tendency to increase the number of ongoing studies of varying complexity and focus, which is directly related to significant progress in various fields of science. As a result, the growth of ongoing research is directly proportional to the growth of scientific papers presented to scientists in the form of scientific articles published in various journals or scientific and educational publications. However, such many scientific papers entails the formation of the problem of finding and selecting materials that are potentially useful for research conducted by young scientists or scientists working in new areas of science. As a solution to this problem, this work is proposed for consideration, devoted to the implementation of its own system designed for natural language processing of scientific publications, to group input data materials into main categories, by extracting attributes from texts using a matrix of terms, reducing the dimension by the Correlation-based approach feature selection and clustering by the k-means algorithm. The result of this solution will be a structured data cascade, consisting of the main topics containing research papers belonging to the input data array. [ABSTRACT FROM AUTHOR]

Techniques for detecting oral cancer are a developing area in medicine that includes cancer diagnosis and early detection. For the classification of oral cancer, we have tried to develop machine learning models like logistic regression (LR), naive Bayes (NB), K-Nearest Neighbors (KNN), support vector machine (SVM), decision tree (DT), and random forest (RF), as well as hybrid models like KNN-SVM and SVM-KNN and deep learning models like Xception, VGG, CNN, Inception, and MobileNet. A few researchers have already used the aforementioned models to categorize oral cancer using genetic data. Nevertheless, the database used for this study includes pictures of the lips and tongue that have been divided into malignant and non-cancerous categories. With the assistance of ENT doctors, the photos were categorized after being captured in various ENT hospitals in Ahmedabad. To determine which model is the most effective, the ML, hybrid, and DL models are evaluated based on numerous evaluation criteria. Experimental result shows that the accuracy, precision, recall, and Fl score of classification of the ML, Hybrid, and deep learning are discussed in this research paper. Out of which hybrid model SVM-KNN with PCA has 74%, Gradient boosting with PCA with 73%, and deep learning model VGG has I 00% accuracy. Hence, it may be concluded that Gradient boosting with PCA and advanced deep learning models like VGG, and Xception is the most favorable model for predicting the survival rate of oral cancer patients by classifying them into cancerous or non-cancerous. This research study motivates towards the early prediction comparatively reduces the severity, and economic cost as the treatment for early diagnosis is less. [ABSTRACT FROM AUTHOR]

Online loans are an attractive financial alternative for people who need funds urgently because they are effortless and efficient. Public opinion about online loans continues to appear on social media platforms, one of which is Twitter. Tweets about online loans gave rise to all kinds of positive and negative opinions. Sentiment analysis is a technique of understanding and identifying opinions from a dataset in text form to produce positive and negative sentiments. In this study, the dataset was taken from tweets on Twitter with the keyword #pinjol and produced 6000 data, consisting of 5677 tweets with negative sentiment and 1230 tweets with positive sentiment. In the classification using the Naïve Bayes algorithm with weighting using the TF-IDF method, this algorithm will be compared with the K-Nearest Neighbors algorithm as a performance comparison in classifying sentiment analysis. The Naïve Bayes method shows a slightly lower accuracy of 99.71% compared to KNN, with an accuracy value of 99.78% at a ratio of 20:80. Meanwhile, in the ratio of 30:70, the Naïve Bayes method shows the best accuracy, which is 99.71% compared to KNN with an accuracy value of 99.61%. [ABSTRACT FROM AUTHOR]

Irrigation is important in plant growth because it controls the need for water. The Binus irrigation system is operated manually, making it difficult to control the water requirements of the plants and thus increasing energy costs. This study explains how machine learning algorithms and the Internet of Things can efficiently control water consumption. This study contributes to the selection of machine learning methods (k-nearest neighbor (KNN), support vector machine (SVM), and Decision Tree (DT)) that can be used in irrigation systems. This system makes use of data from pH sensors, temperature sensors, and soil moisture sensors. The data is divided into 20% training and 80% testing data, and the performance of the three algorithms is compared. The test results show that the SVM algorithm performs better than the other two algorithms in terms of accuracy by 99.35 percent. The hyperplane technique can accurately classify test data despite being trained with a relatively small data set of 206 datasets and two classes. The SVM algorithm is used in the Binus irrigation system to reduce water consumption by 4%. [ABSTRACT FROM AUTHOR]

The world is now a field for perpetual rat race in every sphere of life and it has grown to be a space where about one thirds of the population suffers from anxiety disorder, proving it to be the most prevalent mental disorder. Keeping all concerns and the greatly developing technologies in mind, this study uses an intelligent analytics method for the responses of the DASS 21 Questionnaire, to train a K-Nearest Neighbour(KNN) model to predict psychological disorders and works on increasing the accuracy of the model. Machine learning models are of immense significance to numerous clinical issues as it may be helpful in diagnosing an illness, to analyze its symptoms and to forecast the result of a diagnosis. K-Nearest Neighbors algorithm (KNN) is found to generate the best performance in predictive analysis of this study. Different models of KNN like 1-NN, 3-NN, 5-NN and 7-NN are examined for the model among which 5-NN was found to be the best. Average accuracy, precision, recall and f-score noted with 5-NN were 89.6, 91.3, 0.89, 0.8975 respectively. Also a latency of 4.4 seconds was found with 5-NN. In the process, Emotional process, linguistic style, temporal process were also utilized for different KNN techniques to improve the results and predictions of the model. [ABSTRACT FROM AUTHOR]

Shorea, well known as Meranti, is one of the most important woods in Indonesia. However, identifying the species is still challenging, especially in the timber trade, to prevent misclassification and document falsification. Thus, a non-destructive method is required as a rapid and handy tool in the field. This study aimed to determine three species of Shorea, namely Shorea parvifolia, S. leprosula and S. bracteolata, originating from Indonesia (Sumatra and Kalimantan). The determination of the woods species was used Near-Infrared Spectroscopy (NIRS), and then validated by their anatomical characteristics. For NIRS, the multivariate analysis, namely Principal Component Analysis (PCA) and a pre-treatment Savitzky-Golay (SG) filter were used To collect the NIR spectral signal at wavenumbers of 8000-4000 cm−1. Further, the data was classified using the k-Nearest Neighbor (k-NN) and Support Vector Machine (SVM) classifiers. The results showed that the accuracy of both algorithms was approximately 67%, which is relatively low. k-NN was able to differentiate S. parvifolia with high accuracy but cannot separate S. leprosula and S. bracteolata. While SVM approaches better separation for S. leprosula and S. bracteolata, even though the matrix score was 50:50. The SVM is probably better at classifying because it does not require complicated data processing like k-NN. The observation of anatomical structures revealed differences between Shorea species. Crystals presented in S. parvifolia and S. leprosula; however, radial canals were observed only in S. leprosula. Crystals were absent in S. bracteolata, but silica bodies were found in rays parenchyma. The observation of microscopic structures can differentiate three species of Meranti wood up to the species level, but this method takes time and requires specific knowledge. Thus, NIRS is feasible as an alternative for rapid wood identification. [ABSTRACT FROM AUTHOR]

Due to its outstanding performance despite its sensitivity to input parameters, the bat algorithm (BA) is a popularly used metaheuristic algorithm in numerous applications. The algorithm's convergence behavior is affected by the algorithm's frequency, the number of bats, pulse emission rate, and loudness parameters. These parameters must be set to their optimal values to increase the outcome's quality. This research aims to find the optimal number of bats of the binary bat algorithm (BBA). In this situation, the fitness criterion for the fitness function will be a classifier error rate. K-nearest neighbors (KNN) are used as the classifier for error rate calculation and validation methods. This study is conducted using 13 datasets with various dimensional sizes. The analysis results indicate that the number of bats substantially impacts the classification accuracy. Based on the results, the optimal number of bats must be considered when tuning the algorithm's parameters to achieve adequate performance. [ABSTRACT FROM AUTHOR]

The sine and cosine algorithm has become a widely researched swarm optimization method in recent years due to its simplicity and effectiveness. Based on the advantages, the study in this paper delves deeper into the key parameters that influence the performance of the algorithm, and has implemented modifications such as integrating the reverse learning algorithm and adding elite opposition solution to create the modified Sine and Cosine Algorithm (the modified SCA). Furthermore, by combining the fuzzy k-nearest neighbor method with the modified SCA, the study simulates numeric datasets with two or multiple classes, and analyzes the results. The accuracy rate (ACC) achieved by the modified SCA FKNN in this paper is compared to other models, with data comparison results and tables presented for each. The modified SCA FKNN proposed in this paper has obvious advantages on accuracy rate(ACC). [ABSTRACT FROM AUTHOR]