JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 P-ISSN: 2656-1743 |E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 481 Classification of Batu Bara Songket Using Gray-Level Co-Occurrence Matrix and Support Vector Machine Sriani-1, Muhammad Siddik Hasibuan-2, Rizkika Ananda-3 Department of Computer Science, Faculty of Science and Technology Universitas Islam Negeri (UIN) Sumatera Utara Medan, Indonesia https://ilkomp.uinsu.ac.id/ sriani@uinsu.ac.id, muhammadsiddik@uinsu.ac.id, rizkika.ananda@uinsu.ac.id (*) Corresponding Author Abstract Songket is a traditional woven cloth from the Malay and Minangkabau tribes. Songket can also be classified from the brocade woven family and woven with gold or silver thread. Songket cloth's beauty is the Indonesian people's wealth and preservation. Batu Bara Regency is one of Indonesia's regions with several Songket motifs characteristics. Public knowledge of Batu Bara Songket motifs is still minimal, and the differences between one motif and another are still unknown. This research provides information about the variety of Songket fabrics by classifying six types of Batu Bara Songket motifs, namely the Bunga Tanjung motif, Pucuk Betikam motif, Pucuk Cempaka motif, Pucuk pandan motif, Tampuk Manggis motif and Tolab Berantai motif based on the extraction of the Gray Level texture feature. The Co-Occurrence Matrix includes four parameters: Contrast, Correlation, Energy, and Homogeneity, as well as a classification method with a Support Vector Machine. The feature extraction values process as input for classification using a Support Vector Machine. The highest accuracy achieved in this study was 57%, using 60 training data and 30 test data. Keywords: Classification; Batu Bara Songket Motif; Gray Level Co-Occurrence Matri; Support Vector Machine Abstrak Songket merupakan jenis kain tenunan tradisional yang berasal dari suku melayu dan Minangkabau. Songket juga dapat digolongkan dari keluarga tenunan brokat dan dapat ditenun dengan benang emas dan perak. Keindahan kain songket merupakan kekayaan masyarakat Indonesia yang harus terus dilestarikan. Kabupaten Batu Bara merupakan salah satu wilayah di Indonesia yang memiliki beberapa ciri khas motif kain songket. Pengetahuan masyarakat akan motif-motif songket Batu Bara masih minim dan perbedaan antara motif yang satu dengan motif yang lain masih belum diketahui. Penelitian ini dibuat dengan tujuan untuk memberikan informasi tentang ragam kain songket dengan mengklasifikasi enam jenis motif songket Batu Bara yakni motif bunga tanjung, motif pucuk betikam, motif pucuk cempaka, motif pucuk pandan, motif tampuk manggis dan motif tolab berantai berdasarkan ekstraksi ciri tekstur Grey Level Co-Occurrence Matrix meliputi empat parameter yakni Contrast, Correlation, Energy, dan Homogeneity, serta metode klasifikasi dengan Support Vector Machine. Nilai ektraksi ciri tersebut selanjutnya akan diproses menjadi masukan untuk klasifikasi menggunakan Support Vector Machine. Akurasi tertinggi yang dicapai dalam penelitian ini sebesar 57 %, dengan menggunakan 60 data latih dan 30 data uji. Kata kunci: Klasifikasi; Motif Songket Batu Bara; Grey Level Co-Occurrence; Support Vector Machine INTRODUCTION Indonesia is a country that is rich in unique and distinctive cultural heritage diversity. Every nation or tribe has a culture (Tahrir et al., 2017). A diverse cultural heritage can become essential for Indonesia, and its preservation is mandatory. One manifestation of the results of this cultural process is the creation of works of art that all Indonesian ethnic groups own. Indonesia's diverse traditional fabrics result from cultural processes, geographical differences, flora, fauna, lifestyle differences, and livelihoods producing various traditional fabrics. Indonesian Traditional Fabrics are in great demand P-ISSN: 2656-1743 | E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 482 in national and international markets (Johan Wahyudi & Ihdahubbi Maulida, 2019). One of the ethnic cultures in Indonesia is Malay, especially in North Sumatra (Rigitta, 2021). Songket is a traditional Malay and Minangkabau woven fabric in Indonesia, Malaysia, and Brunei (Nurhalimah et al., 2020). Songket belongs to the brocade woven family. Weaving Songket cloth by hand using gold and silver threads is generally worn on formal occasions. The beauty of Songket can attract domestic and foreign tourists who like Indonesian traditional cloth art (Salamah & Kusumanto, 2017). Batu Bara Regency has a type of Songket with its characteristics (Abdiansyah, 2018), Public knowledge of the Batu Bara Songket motifs is still minimal, and the difference between one motif and another is still unknown. The lack of automated data collection is the cause of this, and no application can analyze the types of Batu Bara Songket motifs, which can help the community to provide knowledge to the public about the Batu Bara Songket motif and is no longer wrong in recognizing the Batu Bara Songket motif. Several studies regarding the classification of Songket motifs based on texture have several times in previous studies, such as the feature extraction study using the grey-level co-occurrence matrix (GLCM) method. The Gabor filter for image classification of Pekalongan batik (Surya et al., 2017), other studies on the classification of Songket cloth in Lombok use GLCM and moment invariant as well as linear discriminant analysis (LDA) (Nurhalimah, 2020), feature extraction of Songket images based on texture using the grey level co- occurrence matrix (GLCM) method (Amalia, 2018), application of a speeded-up robust feature on the random forest for classification of Palembang Songket motifs (Yohannes et al., 2020). This study has advantages over previous research, namely using grounded theory through qualitative analysis using Songket cloth objects that are observed and interacted with based on the participants' views. This research was made to provide information about the various types of coal Songket cloth that previous researchers have not studied by classifying six types of Batu Bara Songket motifs, namely six types of Batu Bara Songket motifs Namely Bunga Tanjung Motif, Pucuk Betikam Motif, Pucuk Cempaka Motif, Pucuk Pandan Motif, Tampuk Manggis Motif, and Tolab Berantai Motif. Gray Level Co-Occurrence Matrix (GLCM) method is a method for extracting image textures. Texture extraction to retrieve essential information from an image before it is used for the following process, using feature extraction methods considered optimal in research (Ramadhani & Bethaningtyas Dyah, 2018). The Gray Level Co- Occurrence Matrix (GLCM) method is an adequate texture descriptor and has better accuracy and computation time than other texture extraction methods (Widodo et al., 2018). The Support Vector Machine (SVM) method is a machine learning technique. They learn by using a pair of input and output data as the desired target. It is called supervised learning, and the advantages of the Support Vector Machine (SVM) method are in recognizing and classifying an object (Anggraini, 2017). Support Vector Machine (SVM) is a classification method with high generalizability and input space dimensions (Neneng et al., 2016). RESEARCH METHODS Types of research The author's research is a type of grounded theory through the analysis of qualitative and quantitative methods. Time and Place of Research This research was conducted precisely at the "Yusra" craftsmen of Batu Bara woven cloth. Padang Genting Village No. 6 district Talawi and the place for system design in the computer laboratory of the Faculty of Science and Technology, State Islamic University of North Sumatra. Research Time in November 2021 to March 2022. Research Target / Subject The research target in this study is the Batu Bara community, who do not know much about the types of Batu Bara songket motifs. Procedure The process carried out to research the classification of Batu Bara Songket motifs based on texture with the Gray Level Co-Occurrence Matrix (GLCM) and Support Vector Machine (SVM) methods, namely through several stages of designing the analytical method. JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 P-ISSN: 2656-1743 |E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 483 Songket Motif RGB Image Da ta Shar ing Training Data Test Data Greysca le Greysca le GLCM Feature Extraction GLCM Feature Extraction Classification Result Analysis SVM Learnin g Outcomes SVM Learnin g Hyperplane Search Training Figure 1. Songket Motif Classification System Planning Diagram Figure 1 shows that the diagram uses input as an RGB Songket motif image. The image is divided into two parts: the training and test data images. Then the training and test data images are changed from RGB to grayscale. After that, the GLCM feature extraction is sought, which consists of contrast, entropy, energy, and homogeneity features. For training data, multiclass SVM learning is used in the hyperplane separator of the six types of Songket motifs. The last stage is to test the system on the test data images and analyze the results of image classification on the hyperplane function that has been obtained. The steps taken in the GLCM calculation are as follows the formation of the initial GLCM matrix from pairs of two parallel pixels corresponding to the directions 0°, 45°, 90°, and 135°. The following form a symmetrical matrix by adding the initial matrix GLCM with its transpose values, normalizing the GLCM matrix by dividing each matrix element by the number of pixel pairs, and then feature extraction, namely contrast, Homogeneity, energy, correlation (Widodo et al., 2018). Contrast = ∑ ∑ (𝑖1 − 𝑖2) 2𝑝(𝑖1, 𝑖2)𝑖2𝑖1 ................................ (1) Homogeneity = ∑ ∑ 𝑝(𝑖1,𝑖2) 1+|𝑖1−𝑖2| 𝑖2𝑖1 ..................................................... (2) Energy = ∑ ∑ 𝑝2(𝑖1, 𝑖2)𝑖2𝑖1 ....................................... (3) Correlation = ∑ ∑ (𝑖−𝜇𝑖)(𝑗−𝜇𝑗)𝑝(𝑖,𝑗) 𝜎𝑖 𝜎𝑗𝑗𝑖 ......................................... (4) In the process of classifying Songket motifs using SVM, in research, in this case, the SVM multiclass approach that It uses is a classification method “one against all.” In this method, k binary SVM models are built, with k being multiple classes. Each classification of the model it wants to use uses total data to find solutions to problems. SVM classifies two classes between one class and others seen as one class. The class for a data sample is directly determined by this method. When the data sample is not included in the group containing the set class but in a specific class, then that class is a class from the sample data in question (Pitoyo, 2020). Table 1. SVM Classification with One-Agains-All method yi = 1 yi=-1 Kernel Hypothesis Class 1 Not class 1 F1(x)=(w1)x+b1 Class 2 Not class 2 F2(x)=(w2)x+b2 .................................................................................................. (5) Table 1 shows test results on test data that produce a decision function with the maximum value given a value of y_i=1 (true), while other decision functions give a value of y_i=-1 (false). Accuracy : Accuracy = 𝐶𝑜𝑟𝑟𝑒𝑐𝑡 𝑐𝑙𝑎𝑠𝑠𝑖𝑓𝑖𝑐𝑎𝑡𝑖𝑜𝑛 𝑛𝑢𝑚𝑏𝑒𝑟 𝐴𝑚𝑜𝑢𝑛𝑡 𝑜𝑓 𝑑𝑎𝑡𝑎 × 100 % ................. (6) Data, Instruments, and Data Collection Techniques Songket motif image data collection is divided into training and test data. The training data consists of 60 samples of Songket motif image data, and the test data consists of 30 samples of Songket motifs. The total sample data is 90 images of the Songket motif. The data collection technique used in this study is : 1. Interview Figure 2. is an interview technique that was conducted by seeking information and knowledge sourced from experts engaged in fields related to this research, namely with Mrs. Hj. Ratna, one of the craftsmen who also opened a Songket business in the Batubara district, so the author gets relevant data references and knows the names of the Batu Bara Songket motifs. P-ISSN: 2656-1743 | E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 484 Figure 2. Interview with Mrs. Hj. Ratna 2. Observation The technique is an observation of data collection techniques by observing directly an object to seek information and knowledge related to research. Image Motif Bunga Tanjung Pucuk Betikam Pucuk Cempaka Image Motif Pucuk Pandan Tampuk Manggis Tolab Berantai Figure 3. Sample image of Batu Bara Songket In Figure 3. It is an example of a sample of Batu Bara Songket images, where each Songket sample shown in Figure 3. has 30 image data. The six types of Songket motifs are Tanjung Flowers, Betikam Shoots, Cempaka Shoots, Pandan Shoots, Tampuk Manggis, and Tolab Berantai. 3. Library Studies, Namely in this study also used literature studies taken from scientific articles, books, and others. Data analysis technique This analysis is needed to determine what kind of software will be produced. The needs analysis in this study is as follows: Start RGB Image Input RGB Image Input Image Transformation Extraction Value Extraction Metho d GLCM End Figure 4. GLCM Feature Extraction Flowchart From the flowchart Figure 4. The above explains the stages in the GLCM feature extraction method, namely as follows : a. Input the RGB image of the Batu Bara Songket motif. b. Then the image is changed from RGB to grayscale. c. After that, look for the GLCM feature extraction, which consists of contrast, Homogeneity, energy, and entropy features. d. Generates image extraction values from the GLCM method. Start Extraction Value Data Normalization Results Data Normalization Results of Determining Output Value Y Determination of Output Value Y End Determination of W eight (W) and Bias (B) Values Result Values ​​(W) and (B) Finding H yperplane as a classification function Classification Results Figure 5. SVM Classification Method Flowchart JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 P-ISSN: 2656-1743 |E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 485 From the flowchart Figure 5. above explains the stages in the Support Vector Machine (SVM) classification method, namely as follows: a. After getting the value from feature extraction From the GLCM method, input the GLCM feature data. b. Data normalization was using the equation formula attached to the SVM theory. c. After Normalization, can pay, pay output value (y). d. Then the value of weight (w) and bias (b). e. After all, can be seen, the system looks for hyperplane as a decision function. f. So the classification uses the SVM method. Output in the Batu Bara Songket motif classification system based on texture using the Gray Level Co-Occurrence Matrix (GLCM) method and Support Vector Machine (SVM), namely the results of the classification of Songket motif types through a feature extraction process using the Gray Level Co-Occurrence method Matrix (GLCM) and the classification process uses the Support Vector Machine (SVM) method. RESULTS AND DISCUSSION Testing Based on the existing image samples, a testing process on these images. At the testing stage, the digital image is in (*.jpg) format with a size of 512x512 pixels. The system testing process using the MATLAB application see in the process below: 1. Application Initial Screen Figure 6. Initial Display Form Figure 6. is the initial form, which is the main page for running the program to be worked on. 2. Image Input Display Figure 7. Image Input Display Form Figure 7. in this form, the image input is by pressing the Image Input button, and then the system will direct it to select the data to be tested, and then the system will automatically display the inputted image and the image file name. 3. Grayscale Display Figure 8. RGB Display Form to Grayscale In figure 8. this form will be processed using a Grayscale. By pressing the grayscale button, the system will process the RGB file to grayscale and display the resulting grayscale image on axes2. 4. Gray Level Co-Occurrence Matrix (GLCM) Feature Extraction Display Figure 9. Gray Level Co-Occurrence Matrix Feature Extract Display Form Figure 9. this form will perform feature extraction on the image by pressing the Feature Extract button. The table will display the feature extraction value of the Gray Level Co-Occurrence Matrix (GLCM). P-ISSN: 2656-1743 | E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 486 5. Classification Result Display Figure 10. Image Classification Results Display Form of Batu Bara Songket Motifs Example 10 The supplied image's classification results show on this form's display. Clicking the classification button will cause the system to display the supplied image's classification findings automatically. System Test Results Based on test data on the image of the type of Songket Batu Bara motif that has been, if there is, in this case, a testing process will be carried out on the motif image Songket Batu Bara with format (*.jpg). In the process of testing the motif classification Songket below, there are 30 test data with 5 data from each type of Songket motif, 5 test data for Bunga Tanjung, 5 test data for Pucuk Betikam, 5 test data for Pucuk Cempaka, 5 test data for Pucuk Pandan, 5 test data for Tampuk Manggis, and 5 test data for Tolab Berantai. From the result testing of as many as 30 test data. The following are the results of system testing of each image of the Songket motif tested to obtain a classification of the type of Coal Songket motif, see table 2 below: Table 2. Data Testing No. Decision Function SVM MultiClass One Against All Score yi Types of Batu Bara Songket Classification Result Information 1. f1(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Bunga Tanjung 'Motif Bunga Tanjung' True 2. f1(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Bunga Tanjung 'Motif Bunga Tanjung' True 3. f1(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Bunga Tanjung 'Motif Bunga Tanjung' True 4. f1(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Bunga Tanjung 'Motif Bunga Tanjung' True 5. f1(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Bunga Tanjung 'Motif Bunga Tanjung' True 6. f2(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) -1 Pucuk Betikam 'Motif Pucuk Pandan' False 7. f2(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Pucuk Betikam 'Motif Pucuk Betikam' True 8. f2(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Pucuk Betikam 'Motif Pucuk Betikam' True …. …………… …. ……….. ……… 30. f6(x)=sign(w1.x1+ w2.x2+ w3.x3+ w4.x4+b) 1 Tolab Berantai 'Motif Tolab Berantai' True JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 P-ISSN: 2656-1743 |E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 487 Table 2 shows 30 test data with 5 data from each type of Songket motif. Each test data has a one against all SVM decision functions according to each class. Entering the yi value shows the classification results and the types of Songket motifs. If the yi value is one, then the classification results are declared correct by the type of Songket motif being tested. Conversely, if the yi value is -1, then the classification results are declared wrong and do not produce output that matches the type of Songket motif in table 3. Table 3. Classification Test Results No Image File Classification Result Average Extraction Result 1. Bunga Tanjung 01.jpg ‘Motif Bunga Tanjung’ Contrast = 0.055182 Correlation = 0.89292 Energy = 0.80746 Homogeneity = 0.97932 2. Bunga Tanjung 02.jpg ‘Motif Bunga Tanjung’ Contrast = 0.054825 Correlation = 0.90229 Energy = 0.89758 Homogeneity = 0.98303 3. Bunga Tanjung 03.jpg 'Motif Bunga Tanjung' Contrast = 0.044537 Correlation = 0.9417 Energy = 0.79726 Homogeneity = 0.98128 4. Bunga Tanjung 04.jpg 'Motif Bunga Tanjung' Contrast = 0.081439 Correlation = 0.81279 Energy = 0.91401 Homogeneity = 0.98792 5. Bunga Tanjung 05.jpg 'Motif Bunga Tanjung' Contrast = 0.029251 Correlation = 0.91165 Energy = 0.76756 Homogeneity = 0.98585 6. Pucuk Betikam 01 .jpg 'Motif Pucuk Pandan' Contrast = 0.59997 Correlation = 0.82928 Energy = 0.091888 Homogeneity = 0.77182 7. Pucuk Betikam 02.jpg 'Motif Pucuk Betikam' Contrast = 0.45013 Correlation = 0.75828 Energy = 0.14618 Homogeneity = 0.81243 8. Pucuk Betikam 03.jpg 'Motif Pucuk Betikam' Contrast = 1.4175 Correlation = 0.67485 Energy = 0.097793 Homogeneity = 0.70073 9. Pucuk Betikam 04.jpg 'Motif Pucuk Pandan' Contrast = 0.4908 Correlation = 0.91867 Energy = 0.10273 Homogeneity = 0.80891 10. Pucuk Betikam 05.jpg 'Motif Pucuk Pandan' Contrast = 0.40894 Correlation = 0.85346 Energy = 0.11702 Homogeneity = 0.82458 11. Pucuk Cempaka 01.jpg 'Motif Pucuk Pandan' Contrast = 0.27883 Correlation = 0.92845 Energy = 0.12754 Homogeneity = 0.86776 P-ISSN: 2656-1743 | E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 488 Continue Table 3. Classification Test Results No Image File Classification Result Average Extraction Result 12. Pucuk Cempaka 02.jpg 'Motif Pucuk Pandan' Contrast = 0.53642 Correlation = 0.92068 Energy = 0.070588 Homogeneity = 0.79759 13. Pucuk Cempaka 03.jpg 'Motif Pucuk Pandan' Contrast = 0.47518 Correlation = 0.93722 Energy = 0.087429 Homogeneity = 0.81312 14. Pucuk Cempaka 04.jpg 'Motif Pucuk Betikam' Contrast = 2.06 Correlation = 0.78755 Energy = 0.060153 Homogeneity = 0.64077 15. Pucuk Cempaka 05.jpg 'Motif Pucuk Pandan' Contrast = 0.26947 Correlation = 0.9625 Energy = 0.13427 Homogeneity = 0.89281 16. Pucuk Pandan 01.jpg 'Motif Pucuk Pandan' Contrast = 0.20786 Correlation = 0.98435 Energy = 0.20333 Homogeneity = 0.91465 17. Pucuk Pandan 02.jpg 'Motif Tolab Berantai' Contrast = 0.7327 Correlation = 0.9424 Energy = 0.32664 Homogeneity = 0.85992 18. Pucuk Pandan 03.jpg 'Motif Pucuk Pandan' Contrast = 0.2666 Correlation = 0.97432 Energy = 0.30293 Homogeneity = 0.90128 19. Pucuk Pandan 04.jpg 'Motif Pucuk Pandan' Contrast = 0.27675 Correlation = 0.97278 Energy = 0.19923 Homogeneity = 0.89126 20. Pucuk Pandan 05.jpg 'Motif Pucuk Pandan' Contrast = 0.21481 Correlation = 0.98068 Energy = 0.25459 Homogeneity = 0.91031 21. Tampuk Manggis 01.jpg 'Motif Tampuk Manggis' Contrast = 0.52105 Correlation = 0.94653 Energy = 0.14224 Homogeneity = 0.85553 22. Tampuk Manggis 02.jpg 'Motif Pucuk Pandan' Contrast = 0.29603 Correlation = 0.95872 Energy = 0.13533 Homogeneity = 0.88979 23. Tampuk Manggis 03.jpg 'Motif Pucuk Pandan' Contrast = 0.56491 Correlation = 0.9307 Energy = 0.28442 Homogeneity = 0.87347 24. Tampuk Manggis 04.jpg 'Motif Pucuk Pandan' Contrast = 0.81957 Correlation = 0.88644 Energy = 0.085617 Homogeneity = 0.7813 JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 P-ISSN: 2656-1743 |E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 489 Continue Table 3. Classification Test Results Table 3 shows 30 test data with 5 data from each type of Songket motif. There is a Songket image file with the file name according to the name of the Songket motif. The classification results are found in the system according to the class of Songket motifs tested. The system's average classification results are obtained according to the class of Songket motifs tested. From the test results of all 30 test data, there are 13 types of Songket motifs that are wrong in the placement of Songket motifs according to their class, so from the results of the classification of Songket motifs, the accuracy results are obtained with a value of 57% with a description of 17 test data that are correct for class placement and 13 incorrect test data in class placement. CONCLUSIONS AND SUGGESTIONS Conclusion Based on the results of tests carried out by classifying the Batu Bara Songket motif based on the image of the Songket motif using the Gray Level Co-Occurrence Matrix (GLCM) and Support Vector Machine (SVM) methods, the following conclusions were drawn: In the manual calculation of the Support Vector Machine (SVM) classification using the one against- all decision function equation with class = sign(f(x)). The function sign is used to check the results of the calculations performed on the test data. The test results on the test data produce the decision with the maximum value given a value of y_i=1, while the other decision functions are given a value of y_i=-1. From the results of testing all test data, which are 30 test data, from the results of the classification of the type of Songket motif, the accuracy results with a value of 57% with a description of 17 test data that is correct in class placement and 13 test data are wrong in class placement. Extraction of Gray Level Co-Occurrence Matrix (GLCM) features of Batu Bara Songket motif images used to classify types of Batu Bara Songket motifs. The distance of neighboring pixels (distance) is one and in the direction of 0°, 45°, 90°, and 135°. The resulting feature of the Gray Level Co- Occurrence Matrix (GLCM) represents the texture value of the Songket motif image. So that these values are used to classify the types of Batu Bara No Image File Classification Result Average Extraction Result 25. Tampuk Manggis 05.jpg 'Motif Pucuk Cempaka' Contrast = 1.7677 Correlation = 0.86206 Energy = 0.064323 Homogeneity = 0.69035 26. Tolab Berantai 01.jpg 'Motif Tolab Berantai' Contrast = 0.33832 Correlation = 0.96599 Energy = 0.46026 Homogeneity = 0.90777 27. Tolab Berantai 02.jpg 'Motif Tolab Berantai' Contrast = 0.82393 Correlation = 0.88715 Energy = 0.46915 Homogeneity = 0.86199 28. Tolab Berantai 03.jpg 'Motif Tolab Berantai' Contrast = 0.62157 Correlation = 0.88225 Energy = 0.46947 Homogeneity = 0.90785 29. Tolab Berantai 04.jpg 'Motif Tolab Berantai' Contrast = 0.24126 Correlation = 0.96976 Energy = 0.47912 Homogeneity = 0.92227 30. Tolab Berantai 05.jpg 'Motif Tolab Berantai' Contrast = 0.37501 Correlation = 0.97167 Energy = 0.46261 Homogeneity = 0.90785 P-ISSN: 2656-1743 | E-ISSN: 2656-1735 DOI: https://doi.org/10.34288/jri.v5i1.469 JURNAL RISET INFORMATIKA Vol. 5, No. 1. December 2022 Accredited rank 3 (SINTA 3), excerpts from the decision of the Minister of RISTEK-BRIN No. 200/M/KPT/2020 490 Songket motifs using the Support Vector Machine (SVM) method, conclusions should be in the form of paragraphs that answer the research objectives. It tells how the researcher’s work can advance current knowledge but does not seem to discuss it. Suggestion The addition of the type of Songket motif studied is universal. The use of different methods as a comparison of this study. It expanded using other feature extraction methods, such as color or shape extraction. Image capture of Songket motifs is idealized in terms of lighting. Moreover, it can be developed into a mobile-based application or website so the wider community can use it. REFERENCES Abdiansyah, M. (2018). Peran Dinas Pendidikan dan Kebudayaan dalam mempromosikan budaya kain tenun songket di desa Padang Genting Kabupaten Batubara. 82. Amalia, I. (2018). Ekstraksi Fitur Citra Songket Berdasarkan Tekstur Menggunakan Metode Gray Level Co-occurrence Matrix (GLCM). Jurnal Infomedia, 3(2), 64–68. https://doi.org/10.30811/jim.v3i2.715 Anggraini, R. (2017). Klasifikasi Jenis Kualitas Keju Dengan Menggunakan Metode Gray Level Co- occurrence Matrix (GLCM) dan Support Vector Machine (SVM) Pada Citra Digital. E- Proceeding of Engineering, 4(2), 2035–2042. Johan Wahyudi, & Ihdahubbi Maulida. (2019). Pengenalan Pola Citra Kain Tradisional Menggunakan Glcm Dan Knn. Jurnal Teknologi Informasi Universitas Lambung Mangkurat (JTIULM), 4(2), 43–48. https://doi.org/10.20527/jtiulm.v4i2.37 Neneng, N., Adi, K., & Isnanto, R. (2016). Support Vector Machine Untuk Klasifikasi Citra Jenis Daging Berdasarkan Tekstur Menggunakan Ekstraksi Ciri Gray Level Co-Occurrence Matrices (GLCM). Jurnal Sistem Informasi Bisnis, 6(1), 1. https://doi.org/10.21456/vol6iss1pp1-10 Nurhalimah. (2020). Klasifikasi Kain Songket Lombok Menggunakan Glcm Dan Moment Invariant Serta Linear Discriminant Analysis (Lda). Nurhalimah, N., Suta Wijaya, I. G. P., & Bimantoro, F. (2020). Klasifikasi Kain Songket Lombok Berdasarkan Fitur GLCM dan Moment Invariant Dengan Teknik Pengklasifikasian Linear Discriminant Analysis (LDA). Jurnal Teknologi Informasi, Komputer, Dan Aplikasinya (JTIKA ), 2(2), 173–183. https://doi.org/10.29303/jtika.v2i2.98 Pitoyo, H. A. (2020). Implementasi Metode Support Vector. Ramadhani, M., & Bethaningtyas Dyah, H. K. (2018). Klasifikasi Jenis Jerawat Berdasarkan Tekstur Dengan Menggunakan Metode Glcm Acne Classification Based on Texture Using Glcm Method. E-Proceeding of Engineering, 5(1), 870–876. Rigitta, P. (2021). Makna Tradisi Lisan Dalam Motif Songket Melayu Langkat. Desember, 14(2), 1979–5408. Salamah, I.-, & Kusumanto, R. (2017). Faktor-Faktor Pemanfaatan Teknologi Informasi UKM Kain Tenun Songket Palembang. Jurnal RESTI (Rekayasa Sistem Dan Teknologi Informasi), 1(3), 177–182. https://doi.org/10.29207/resti.v1i3.71 Surya, R. A., Fadlil, A., & Yudhana, A. (2017). Ekstraksi Ciri Metode Gray Level Co- Occurrence Matrix ( GLCM ) dan Filter Gabor untuk Klasifikasi Citra Batik Pekalongan. Jurnal Informatika:Jurnal Pengembangan IT (JPIT , Vol. 02, No. 02, Juli 2017, 02(02), 23–26. Tahrir, R., Rohidi, T. R. &, & Iswidayati, S. (2017). Makna Simbolis dan Fungsi Tenun Songket Bermotif Naga Pada Masyaratakat Melayu di Palembang Sumatera Selatan. Catharsis: Journal of Arts Education, 6(1), 9–18. Widodo, R., Widodo, A. W., & Supriyanto, A. (2018). Pemanfaatan Ciri Gray Level Co-Occurrence Matrix (GLCM) Citra Buah Jeruk Keprok (Citrus reticulata Blanco) untuk Klasifikasi Mutu. Jurnal Pengembangan Teknologi Informasi Dan Ilmu Komputer, 2(11), 5769– 5776. https://j-ptiik.ub.ac.id/index.php/j- ptiik/article/view/3420 Yohannes, Y., Devella, S., & Pandrean, A. H. (2020). Penerapan Speeded-Up Robust Feature pada Random Forest Untuk Klasifikasi Motif Songket Palembang. Jurnal Teknik Informatika Dan Sistem Informasi, 5(3), 360– 369. https://doi.org/10.28932/jutisi.v5i3.1978