SHM Publisher Journals
Not a member yet
315 research outputs found
Sort by
Improved Accuracy of Naive Bayes Classifier for Determination of Customer Churn Uses SMOTE and Genetic Algorithms
Full text linkWith increasing competition in the business world, many companies use data mining techniques to determine the level of customer loyalty. The customer data used in this study is the german credit dataset obtained from UCI. Such data have an imbalance problem of class because the amount of data in the loyal class is more than in the churn class. In addition, there are some irrelevant attributes for customer classification, so attributes selection is needed to get more accurate classification results. One classification algorithm is naive bayes. Naive Bayes has been used as an effective classification for years because it is easy to build and give an independent attribute into its structure. The purpose of this study is to improve the accuracy of the Naive Bayes for customer classification. SMOTE and genetic algorithm do for improving the accuracy. The SMOTE is used to handle class imbalance problems, while the genetic algorithm is used for attributes selection. Accuracy using the Naive Bayes is 47.10%, while the mean accuracy results obtained from the Naive Bayes with the application of the SMOTE is 78.15% and the accuracy obtained from the Naive Bayes with the application of the SMOTE and genetic algorithm is 78.46%
Support Vector Machine (SVM) Optimization Using Grid Search and Unigram to Improve E-Commerce Review Accuracy
Full text linkElectronic Commerce (E-Commerce) is distributing, buying, selling, and marketing goods and services over electronic systems such as the Internet, television, websites, and other computer networks. E-commerce platforms such as amazon.com and Lazada.co.id offer products with various price and quality. Sentiment analysis used to understand the product’s popularity based on customers’ reviews. There are some approaches in sentiment analysis including machine learning. The part of machine learning that focuses on text processing called text mining. One of the techniques in text mining is classification and Support Vector Machine (SVM) is one of the frequently used algorithms to perform classification. Feature and parameter selection in SVM significantly affecting the classification accuracy. In this study, we chose unigram as the feature extraction and grid search as parameter optimization to improve SVM classification accuracy. Two customer review datasets with different language are used which is Amazon reviews that written in English and Lazada reviews in the Indonesian language. 10-folds cross validation and confusion matrix are used to evaluating the experiment results. The experiment results show that applying unigram and grid search on SVM algorithm can improve Amazon review accuracy by 26,4% and Lazada reviews by 4,26%
Improving Algorithm Accuracy K-Nearest Neighbor Using Z-Score Normalization and Particle Swarm Optimization to Predict Customer Churn
Full text linkDue to increased competition in the business world, many companies use data mining techniques to determine the loyalty level of customers. In this business, data mining can be used to determine the loyalty level of customers. Data mining consists of several research models, one of which is classification. One of the most commonly used methods in classification is the K-Nearest Neighbor algorithm. In this study, the data which used are from German Credit Datasets obtained from UCI machine learning repository. The purpose of this study is to find out how Z-Score works to normalize the data and Particle Swarm Optimization to find the most optimal K value parameters, so the performance of the K-Nearest Neighbor algorithm is more optimal during the classification. The methods which were used to normalize the data are Z-score and Particle Swarm Optimization to determine the most optimal K value. The classification was tested using confusion matrix to determine the generated accuracy. From the finding of this study, the application of Z-score normalization and Particle Swarm Optimization with the K Nearest Neighbor algorithm succeed in increasing the accuracy up to 14%. The initial accuracy was 68.5%, and after applying the normalization of Z-Score and Particle Swarm Optimization, the accuracy became 82.5%
Increasing Accuracy of C4.5 Algorithm Using Information Gain Ratio and Adaboost for Classification of Chronic Kidney Disease
Full text linkData information that has been available is very much and will require a very long time to process large amounts of information data. Therefore, data mining is used to process large amounts of data. Data mining methods can be used to classify patient diseases, one of them is chronic kidney disease. This research used the classification tree method classification with the C4.5 algorithm. In the pre-processing process, a feature selection was applied to reduce attributes that did not increase the results of classification accuracy. The feature selection used the gain ratio. The Ensemble method used adaboost, which well known as boosting. The datasets used by Chronic Kidney Dataset (CKD) were obtained from the UCI repository of learning machine. The purpose of this research was applying the information gain ratio and adaboost ensemble to the chronic kidney disease dataset using the C4.5 algorithm and finding out the results of the accuracy of the C4.5 algorithm based on information gain ratio and adaboost ensemble. The results obtained for the default iteration in adaboost which was 50 iterations. The accuracy of C4.5 stand-alone was obtained 96.66%. The accuracy for C4.5 using information gain ratio was obtained 97.5%, while C4.5 method using information gain ratio and adaboost was obtained 98.33%
Data Security System of Text Messaging Based on Android Mobile Devices Using Advanced Encrytion Standard Dynamic S-BOX
Full text linkMost of the recent technologies are turning to mobile platforms, Android becames one of the most widely used OS. Eventhough it has complete features, even it's not safe enough such like Chat Messenger. The security of messages distribution is a challenge to increase of vulnerable distribution of information through the network today. Therefore, a data security or cryptographic algorithm is needed to secure the messages so that it cannot be read by irresponsible people. National Institute of Standard and Technology (NIST) established the Advanced Encrytion Standard (AES) cryptographic algorithm as a standard encryption algorithm that is safe and can be used globally. AES algorithm is included in block cipher cryptography that uses substitution boxes (S-BOX) in its operations, so that algorithmically can make input and output unrelated. So, it can provide more varied output in the process, we need a dynamic S-BOX. In this research, dynamic S-BOX generalized using XOR operations from affine transformations with 8-bit binary element matrices arranged and randomly to produce as many as 256 S-Boxes. The application of dynamic AES with S-BOX algorithm on Android-based messenger chat application is built using the Java programming language and database hierarchy for data storage. The implementation results showed that the algorithm was running well and could encrypt the text of the message to ciphertext and decrypt the ciphertext to the original message. This research can be used as a reference so that further researchers can merge the AES algorithm with other algorithms to improve the security of encryption in text files, documents, images, videos or other types of files