Confusion Matrix Method Research Articles

Enhancing textile quality control with the application of teachable machine and Raspberry Pi as machine learning-based image processing

The adoption of image processing-based technologies in the textile sector is rising. This technology is commonly utilized to replace traditional sensor systems that are limited to a single function while also improving product quality control functions. Defects during the manufacturing process are a common problem in the textile business, particularly with fabric products. This study created a fabric quality control system that detects fabric problems using machine learning-based picture classification techniques. A D320p web camera detects rare and slap flaws, which are classified using open-source Google teaching machine software and processed on a Raspberry Pi 3B device. The laboratory-scale measurement was carried out on a prototype cloth rolling machine using the confusion matrix method. The test results reveal an average inference speed of 143.5 milliseconds, a frame rate of 6.45 fps, and a 98.56% accuracy rate. These results demonstrate that the proposed system is effective and efficient for detecting fabric defects, offering a promising solution for enhancing quality control in the textile industry. Future research could focus on scaling the system for industrial use and enhancing real-time performance.

Sistem Pendukung Keputusan Pemilihan Objek Wisata Kabupaten Berau Menggunakan Metode Additive Ration Assessment (ARAS)

Indonesia is a country that is famous for its natural wealth and tourist destinations worldwide, and it is spread out in various regions. One of the regions in East Kalimantan that is popular and often visited by tourists every year because it is famous for its stunning underwater tourist destinations is the Berau Regency. According to the East Kalimantan Tourism Office, this district has the highest number of tourist attractions in East Kalimantan, which is 224. Many tourist attractions trigger confusion among tourists when deciding which alternative option to visit. Therefore, this research aims to build a decision support system to assist tourists in getting alternative tourism recommendations that are appropriate and to their needs. The method used in this research is the Additive Ratio Assessment (ARAS) method. There are 30 alternative attractions and 5 criteria: facilities, safety, cleanliness, cost, and distance. The results show that the ARAS method can recommend the best alternative to tourist attractions. The accuracy testing results using the confusion matrix method get a percentage value of 80%.

Implementation of Raita Algorithm in Manado-Indonesia Translation Application with Text Suggestion Using Levenshtein Distance Algorithm

Abstract. Manado City is one of the multidimensional and multicultural cities, possessing assets that are considered highly potential for development into tourism and development attractions. The current tourism assets being developed by the Manado City government are cultural tourism, as they hold a charm and allure for tourists. Hence, a communication tool in the form of a translation application is necessary for facilitating communication between visiting tourists and the native community of North Sulawesi, even for newcomers who intend to reside in North Sulawesi, given that the Manado language serves as the primary communication tool within the community. This research employs a combination of the Raita algorithm and the Levenshtein distance algorithm in its creation process, along with the confusion matrix method to calculate the accuracy of translation results using the Levenshtein distance algorithm with a text suggestion feature. The research begins by collecting a dataset consisting of Manado language vocabulary and their translations in Indonesia language, sourced from literature studies and original respondents from North Sulawesi, which have been validated by a validator to prevent translation data errors. The subsequent stage involves preprocessing the dataset, converting the entire content of the dataset to lowercase using the case folding process, and removing spaces at the start and end of texts using the trim function. Next, both algorithms are implemented, with the Raita algorithm serving for translation and the Levenshtein distance algorithm providing text suggestions for typing errors during the translation process. The accuracy results derived from the confusion matrix calculations during the translation process of 100 vocabulary words, accounting for typing errors, indicate that the Levenshtein distance algorithm is capable of effectively translating vocabulary accurately and correctly, even in the presence of typing errors, resulting in a high accuracy rate of 94,17%. Purpose: To determine the implementation of the Levenshtein distance and Raita algorithms in the process of using the Manado-Indonesian translation application, as well as the resulting accuracy level. Methods/Study design/approach: In this study, a combination of the Raita and Levenshtein distance algorithms is utilized in the translation application system, along with the confusion matrix method to calculate accuracy. Result/Findings: The accuracy achieved in the translation process using text suggestions from the Levenshtein distance algorithm is 94.17%. Novelty/Originality/Value: This research demonstrates that the combination of the Raita and Levenshtein distance algorithms yields optimal results in the vocabulary translation process and provides accurate outcomes from the use of effective text suggestions. This is attributed to the fact that nearly all the data used was successfully translated by the system, even in the presence of typographical errors.

Spatiotemporal analysis of land use and land cover changes, LST and NDVI in Thatta district, Sindh, Pakistan

The purpose of this work is to determine land-use and land-cover (LULC) patterns, land surface temperature (LST), and normalized difference vegetation index (NDVI) changes in Thatta district using Landsat data from 1991 to 2021 and evaluate the relationship between LST and NDVI. The research process employed the selection of the study area, data acquisition, preprocessing, and classification of remotely sensed images for the estimation of the land use land cover change (LULC), vegetation index (NDVI), and evaluation of LST using thermal bands in the Landsat dataset. The study revealed the area under built-up structures has increased from 1991 to 2021. Although the vegetation cover showed an increase, the bare soil showed a decreasing pattern, indicating a constant change in the LULC patterns in the region. The confusion matrix method for accuracy valuation of LULC data of 2021 revealed an overall accuracy of 88.24%, with a Kappa coefficient of 84.22%, while the Artificial Neural Network Multilayer Perceptron (ANN-MLP) model had a Kappa validation of 0.95 for 2021. The highest maximum temperature is observed for 2021, indicating a positive relationship between LST and built-up structures, while regression analysis found a negative correlation between LST and NDVI. This study provides a valuable monitoring framework to help resource managers develop strategies to manage land resources.

A VIKOR-Based Decision Support System for Prioritizing Public Facility Improvements in Malang City with Geotagging Integration

Public facilities play a crucial role in driving economic growth and development. Nevertheless, the dearth of public information concerning facility enhancements fosters a sense of public distrust towards the government. Additionally, numerous facilities, which should be prioritized for improvement, have not received adequate attention. In contrast to several prior studies, the present study encompasses a broader scope and incorporates geotagging techniques to precisely identify the location of complaints and determine the optimal route to reach them. Moreover, an analysis process utilizing the VIKOR method has been devised to assess the priority of public facility improvements. This method yielded an accuracy rate of 89,7%, signifying a commendable level of precision and a 16% increase in accuracy based on confusion matrix method compared to previous studies. Through user usability testing, it was determined that the majority of users agreed that this system can facilitate public reporting, enable progress monitoring of public facility improvements, and aid in prioritizing such improvements.

The Influence of Concrete Age on the Accuracy of Pile Integrity Testing on Bored Piles Foundation

Abstract Bored pile foundations are defined as deep foundations in the form of an elongated tube with a specific diameter made of reinforced concrete installed using the drilling method and in-situ casting concrete. The cast in-situ process in bored holes with a deep elevation below the ground surface can potentially cause integrity problems, such as voids, segregation, fracture, necking, bulging, and other concrete defects. Currently, many non-destructive test methods are available to ensure the quality of the bored pile; one is the Low Strain Integrity Testing method, also known as Pile Integrity Testing (PIT). ASTM recommends that PIT should be carried out no sooner than 7 days after casting or the strength of the concrete has reached a minimum of 75,0% of its design strength. In fact, the PIT test is carried out before the concrete’s age and/or its strength is achieved, considering the limited construction time. This could lead to misinterpretation of the PIT test results and inaccurate identification of the bored pile defect locations. The study aims to determine the effect of concrete age on the accuracy of bored pile defect locations on a laboratory scale. The test model consisted of four bored piles with a diameter of 0,3 meters and a depth of 2,0 meters, which were cast using reinforced concrete with a design strength of fc’ 30 MPa. The PIT was performed at <7, 14, and 28 days of concrete age, and then the confusion matrix method was used to create a binary classification model to determine the accuracy of the defect location. The defect locations were visually observed and validated by dismantling the test model after 28 days of casting. The results showed that the accuracy of identifying the location of concrete defects was not affected by the age of the concrete but was determined by the compressive strength of the concrete during testing. Referring to the F-score, bored piles with concrete strength >75,0% of the design accurately identify defect locations above 84,21%. On the other hand, bored piles with concrete strength <75,0% design resulted in accuracy below 66,67%.

Named Entity Recognition In Electronic Medical Records Based On Hybrid Neural Network And Transformer

The development of artificial intelligence in the field of health encourages the use of electronic medical records in all health facilities to record health services provided to patients. For hospitals, extracting information from electronic medical records can make it easier for management to make clinical decisions and for researchers to obtain data for research in the medical and nursing fields. The research builds a model of named entity recognition in electronic medical records based on hybrid neural networks, bidirectional encoder representations from transformers, and setting hyperparameters to get the highest accuracy. The research data set is processed from an initial examination form of an adult patient in a hospital onto an electronic medical record in 2022, and the data is pre-processed. Next, perform the entity-tagging phase on the text and divide 70% of the training datasets by 30% of the testing datasets. Training and evaluation of models built using the confusion matrix method. The results of this study show that the entity identification model called bidirectional encoder representations from transformers consistently outperforms the neural network-based entity recognition model in any evaluation metric. The abbreviation of the bi-directional encoder representation of transformer has very high precision, recall, and f1-score values, demonstrating its ability to recognise entities very well. In this study, although the model named Entity Recognition based on neural networks also has high accuracy, the low precision and recall values indicate that this model may have difficulty recognising entities accurately.

Performance Analysis of the Support Vector Machine Algorithm in Predicting Rain Potential in DKI Jakarta

Indonesia is a country with a tropical climate which is located on the equator which tends to get sunlight throughout the year. Not only gives beauty but also saves disasters that can be dangerous such as floods, which usually occur due to high rainfall. The impact is also large on facilities with damage to buildings and health problems. It is very important to prepare it so that the possibility of damage or loss can be minimized. This research will apply the Support Vector Machine (SVM) Algorithm by selecting the distribution ratio of training and test data as well as the best kernel function to predict the potential for rain using daily climate data from the Meteorology, Climatology and Geophysics Agency (BMKG) in DKI Jakarta with the help of Rstudio software. The performance evaluated using the confusion matrix method produces the highest accuracy value of 89% is the SVM model with a training data distribution ratio of 90% and the Linear kernel as the chosen model for predicting rain potential.

Computer model for detecting tsunami wave hazard on built-up land using machine learning and sentinel 2A satellite imagery

<p>The aim of this research is to compile a tsunami wave hazard scale based on built-up land density extracted and classified by machine learning from Sentinel 2A satellite and digital elevation model (DEM) imageries. This research was carried out in 5 stages, namely: (i) pre-processing of Sentinel 2A and DEM images, (ii) Classification of VI data using the machine learning algorithms, (iii) Spatial prediction using the ordinary kriging method, (iv) Field testing using the confusion matrix method, (v) Preparation of decision matrix for tsunami wave hazard. The results of the study show that the most accurate classification algorithm for classifying built-up indices data is the k-nearest neighbor (k-NN) algorithm. The results of the statistical accuracy test show that the most accurate is normalized difference built-up index (NDBI) with a mean of square error (MSE) value of 0.073 and a mean of absolute error (MAE) of 0.003. DEM analysis shows that the research area is at an altitude of 0–15 meters above sea level so it is in the high vulnerability to medium vulnerability category. Field testing showed user accuracy of 91.11%, manufacturer accuracy of 92.16%, and overall average accuracy of 91%.</p>

Classification of Periapical and Bitewing Radiographs as Periodontally Healthy or Diseased by Deep Learning Algorithms.

Objectives The aim of this artificial intelligence (AI) study was to develop a deep learning algorithm capable of automatically classifying periapical and bitewing radiography images as either periodontally healthy or unhealthy and to assess the algorithm's diagnostic success. Materials and methods The sample of the study consisted of 1120 periapical radiographs (560 periodontally healthy, 560 periodontally unhealthy) and 1498 bitewing radiographs (749 periodontally healthy, 749 periodontally ill). From the main datasets of both radiography types, three sub-datasets were randomly created: a training set (80%), a validation set (10%), and a test set (10%). Using these sub-datasets, a deep learning algorithm was developed with the YOLOv8-cls model (Ultralytics, Los Angeles, California, United States) and trained over 300 epochs. The success of the developed algorithm was evaluated using the confusion matrix method. Results The AI algorithm achieved classification accuracies of 75% or higher for both radiograph types. For bitewing radiographs, the sensitivity, specificity, precision, accuracy, and F1 score values were 0.8243, 0.7162, 0.7439, 0.7703, and 0.7821, respectively. For periapical radiographs, the sensitivity, specificity, precision, accuracy, and F1 score were 0.7500, 0.7500, 0.7500, 0.7500, and 0.7500, respectively. Conclusion The AI models developed in this study demonstrated considerable success in classifying periodontal disease.Future applications may involve employing AI algorithms for assessing periodontal status across various types of radiography images and for automated disease detection.

Comparison and Evaluation of Five Global Land Cover Products on the Tibetan Plateau

The Tibetan Plateau (TP) region contains maximal alpine grassland ecology at the mid-latitudes. This region is also recognized as an ecologically fragile and sensitive area under the effects of global warming. Regional climate modeling and ecosystem research depend on accurate land cover (LC) information. In order to obtain accurate LC information over the TP, the reliability and precision of five moderate/high-resolution LC products (MCD12Q1, C3S-LC, GlobeLand30, GLC_FCS30, and ESA2020 in 2020) were analyzed and evaluated in this study. The different LC products were compared with each other in terms of areal/spatial consistency and assessed with four reference sample datasets (Geo-Wiki, GLCVSS, GOFC-GOLD, and USGS) using the confusion matrix method for accuracy evaluation over the TP. Based on the paired comparison of these five LC datasets, all five LC products show that grass is the major land cover type on the TP, but the range of grass coverage identified by the different products varies noticeably, from 43.35% to 65.49%. The fully consistent spatial regions account for 43.72% of the entire region of the TP, while, in the transition area between grass and bare soil, there is still a large area of medium-to-low consistency. In addition, a comparison of LC datasets using integrated reference datasets shows that the overall accuracies of MCD12Q1, C3S-LC, GlobeLand30, GLC_FCS30, and ESA2020 are 54.29%, 49.32%, 53.03%, 53.73%, and 60.11%, respectively. The producer accuracy of the five products is highest for grass, while glaciers have the most reliable and accurate characteristics among all LC products for users. These findings provide valuable insights for the selection of rational and appropriate LC datasets for studying land-atmosphere interactions and promoting ecological preservation in the TP.

SISTEM PENDUKUNG KEPUTUSAN PENENTUAN STATUS GIZI BALITA PADA DINAS KESEHATAN KABUPATEN ALOR MENGGUNAKAN ALGORITMA C4.5

Assessment of nutritional status plays a very important role in monitoring children's nutrition. If the assessment of nutritional status is carried out correctly and accurately, signs or symptoms of impaired growth and development in children can be identified early so that mitigation and prevention can be carried out in an effort to improve the nutritional status of children under five. And the occurrence of nutritional problems can be minimized. For this reason, determining the nutritional status of toddlers must be done quickly and accurately. Puskesmas as the technical implementer of the Health Service, has the main task of collecting data and assessing the nutritional status of toddlers and submitting the results of the assessment to the Health Service. The determination of nutritional status that has been carried out is by looking at and calculating the standard anthropometric table for assessing the nutritional status of toddlers according to the Decree. Minister of Health Number: 1995/Menkes/SK/XII/2010. However, Puskesmas are often slow in treating large numbers of patients and also have limited resources, especially in areas where there is a shortage of medical personnel. So this also influences the results of the assessment of determining the nutritional status of toddlers. For this reason, the Health Service needs to provide an optimal method for determining the nutritional status of toddlers to be used by Community Health Centers to determine the nutritional status of toddlers quickly, precisely and accurately. So researchers have conducted research and produced a decision support system for determining nutritional status in toddlers using the C4.5 algorithm. Test results to measure algorithm performance using the Confusion Matrix, Accuracy, Precision and Recall testing method, it is known that the C4.5 algorithm has an accuracy value of 72.13 %, Precision value 71.43% and Recall value 22.73% Keywords : Decision Support, Nutritional Status, Toddlers, C4.5 Algorithm

Mapping winter fallow arable lands in Southern China by using a multi-temporal overlapped area minimization threshold method

ABSTRACT In China, a nation facing farmland scarcity, accurate mapping of winter fallow arable lands is crucial for enhancing crop rotation and land use efficiency. The Dynamic Threshold (DT) method commonly used in phenology and winter fallow land studies often employs empirically set subjective thresholds. This approach tends to uniformly apply a single threshold across diverse vegetation types and geographical regions, neglecting the variations in physiological traits and spatial heterogeneity. This study developed a multitemporal overlapping area minimization threshold method (MOAMT). This approach constructed Normalized difference vegetation index (NDVI) probability density functions for both winter fallow and non-winter fallow arable land across a time series to attain a statistically significant threshold. We use MOAMT and DT methods for winter fallow arable land extraction in southern China, respectively. And then, comparison of identification accuracy between MOAMT and DT has been implemented by using confusion matrix method. Compared to the DT method, MOAMT exhibits better performance. The mapping of winter fallow arable lands highlighted their predominant distribution in the middle and lower reaches of the Yangtze River (MLR-YR) Basin, covering approximately 20.8 million hectares, with potential for development. This study will provide information support for the optimization of planting layouts in China, offering new opportunities for further increasing grain production. While in the sense of research approach, this study indicates the superiority of threshold obtained through statistical probability calibration over empirical universal thresholds for classification accuracy. We advocate for prioritizing threshold calibration in future applications of threshold classification methods.

Deteksi Helm Keselamatan Menggunakan Jetson Nano dan YOLOv7

Increasing awareness of the importance of head personal protective equipment in industrial and construction environments has become a major concern in efforts to improve occupational safety. This research developed an early detection system for the use of computer vision-based head protective equipment using the YOLOv7 model and the Jetson Nano controller. The YOLOv7 algorithm was chosen for its ability for fast and accurate object detection. The YOLOv7 model was trained with a total dataset of 2799 images and iterations of 100 epochs to detect head personal protective equipment with a high degree of accuracy. The system captures imagery, activates a warning alarm, and sends a notification to Telegram when a violation occurs on an object that is not wearing a safety helmet. The test results using the confusion matrix method showed that the developed system was able to detect head personal protective equipment with an accuracy rate of 97.23%, which shows the system's ability to recognize personal protective equipment with very high accuracy. In addition, the system also showed a precision value of 98.71% indicating that all detections performed were correct, and a recall of 95.63% which describes the system's ability to recognize most of the head personal protective equipment available. The average FPS result using GPU with CUDA on Jetson Nano reached 5,723 FPS.

Klasifikasi Penyakit Daun Tomat Menggunakan Algoritma K-NN Berdasarkan Ekstraksi Fitur GLCM dan LBP

Tomatoes are widely cultivated in Indonesia and are one of the crops with high economic value. According to the Central Bureau of Statistics, the Indonesian nation has the capacity to produce up to 1. 17 million tons of tomatoes in 2020. Tomatoes contain ingredients that the body needs. Additionally, grapes can also be consumed in different forms. However, tomato production in Indonesia decreased from 2013 to 2015 due to the spread of pests. Therefore, we conducted a study to classify tomato leaf diseases using the K-nearest neighbor method based on the grayscale coexistence matrix and the extraction of local binary pattern features. The data used was 9157 data obtained from the Plant Village database and classified into 6 classes (healthy, spot fungus, late blight, leaf mold, mosaic virus, spider mite, and target spot). The testing process was performed using the K-fold cross-validation technique, followed by performance calculations using the confusion matrix method. The highest accuracy was obtained at 86.8% when classifying using K9 and K10 with a precision value 40.6% and recall 49.2% when classifying using a value of K = 9, and precision 49.7% and recall of 49.3% when classifying using a value of K = 10.

Corporate Financial Risk Identification and Operation Control Analysis for XGBoost Modeling

Abstract Risks in the financial market are omnipresent, and the operations of listed companies are affected by various factors, so the study of the financial risks of listed companies is also of great significance. In this paper, the statement data of listed companies and the text data of annual reports are used separately. The XGBoost model is used to analyze its classification effect, and the confusion matrix and ROC curve evaluation methods are used to compare the accuracy of the prediction results between the XGBoost model and the GBDT model, which helps corporate managers to identify the financial risks of enterprises in advance, and at the same time, improves the level of operation control. The results show that the accuracy of the XGBoost model fluctuates around 0.85, and the highest accuracy of the model is 0.883 when the number of its features is 21. The results of the confusion matrix assessment show that the accuracy of the prediction results of the risk-free company of the XGBoost model reaches 94.95%, and the accuracy of the prediction results of the XGBoost model increases by 5.15% compared with that of the GBDT model. This is in accordance with the ROC curve evaluation results. Obviously, the XGBoost model has a better prediction effect and a more stable early warning performance, and the use of the XGBoost model can help the managers of listed companies to be informed of the deterioration of the company’s financial situation as early as possible so that they can implement the corresponding operational control measures to reduce losses in time.

Performance Comparison of the Combination of Smarter-Fuzzy and Smarter-Fuzzy-Topsis Methods on the accuracy of the PPA Scholarship Acceptance Decision Support System

The PPA Scholarship is a scholarship intended for students who excel, are active in campus activities, and are economically disadvantaged in accordance with predetermined conditions. Currently, STAHN MPU Kuturan Singaraja in calculations often takes a long time, causing delays from the predetermined schedule and less accurate results. Based on this, it is necessary to develop a Decision Support System that can provide recommendations for prospective PPA scholarship recipients. There are 6 criteria being used, including: GPA, semester, credits, charter/certificate, assignment letter, and parent's income. In this study, comparing the performance of the combination of the Smarter-Fuzzy and Smarter-Fuzzy-Topsis methods through testing using the confusion matrix method. Smarter method is used to be able to give weighting criteria with ROC technique. The Fuzzy method is used to convert the criteria values ??for each alternative from 0 to 1. The Topsis method is used to rank. The ranking results of the combination Smarter-Fuzzy method show an accuracy rate of 86.6%, and the combination of the Smarter-Fuzzy-Topsis method of 75.5%.

THE USE OF A MAGNETIC SENSOR TO DETECT LOBSTER (Panulirus spp.) CATCHES ON THE LABORATORY SCALE

Problems in operating trap fish tool are that fishermen do not know the actual condition of the traps. Thus, it causes the effectiveness of time and number of catches to be not optimal. Based on the conditions and problems above, the writer concluded that an innovation in the operation of the trap to design a magnetic sensor-based system/tool to detect lobster catches that can be accessed via smartphone is required. The aim of this study is to design and observe the performance of using a magnetic sensor to detect lobsters' (Panulirus spp.) movements on a laboratory scale (in the aquarium). Data analysis in this study used the Confusion Matrix method, where this method divides the test results into 4 common conditions: TP (True Positive), TN (True Negative), FP (False Positive), and FN (False Negative). From the 4 conditions, Recall, Specificity, Precision, Accuracy, and F1 Score can be calculated. From 16 tests with the 16 lobsters of 60-190 grams/individual, it obtained conditions of TP of 25 times, TN of 109 times, FP of 11 times, and FN of 15 times. From the four conditions, it was obtained a Recall value of 0.625 or 62.50%, Specificity of 0.9083 or 90.83%, Precision of 0.6944 or 69.44%, Accuracy of 0.8375 or 83.75%, and F1 Score of 0.6579 or 65.79%. Based on the observations and test results of the actual detection and application system in this study, performance reference for using magnetic sensors was by using Accuracy with a score of 83.75%.

Text-to-speech on health monitoring bracelet for the visually impaired

Text-to-speech (TTS) is a technology that converts text into sound using a phonetization system and is especially useful to be applied to blind aids who need information in the form of sound because of their limitations. To help the visually impaired know their health conditions where the visually impaired feel limited to going out during the pandemic. For this reason, it is necessary to make an aid that can read text-based data such as body temperature, heart rate per minute, and oxygen levels into a voice that can be heard by the blind, the method used in this study is finite state automata (FSA) which is used to split Indonesian words into words according to its syllable patterns and facilitate the pronunciation process which is included in the blind aids so that it is expected to help the visually impaired to be able to find out their health condition. In this study, the test was carried out using the confusion matrix method and the results obtained were 100% accurate, 99.71% accuracy of temperature sensor, 98% accuracy of heart rate, and 95% accuracy of oxygen saturation.

PERANCANGAN APLIKASI KLASIFIKASI SENTIMEN BERBASIS WEB TERHADAP MATA UANG KRIPTO

Sentiment analysis is a collection of opinions that studies opinions about something like an issue, product, event or topic. In deciding to choose cryptocurrency, the majority of Indonesian people tend to only read the latest reviews about cryptocurrencies and even a few will lead to wrong conclusions. To analyze cryptocurrency sentiment easily and quickly, a sentiment analysis system with a convenient and easy-to-use interface is required. The application to analyze this sentiment is web-based which is made using the PHP (Hypertext Preprocessor) framework by applying the MVC (Model View Controller) concept. This application can import data in CSV format which is then processed into an application using the nave Bayes algorithm in the sentiment classification process which is then evaluated by the confusion matrix method. The reviews are taken from social media twitter and it is hoped that this application will be able to provide classification information accurately, precisely and reliably.