Rapid Miner Tool Research Articles

Segmenting e-Commerce Customer through Data Mining Techniques

Business-to-business e-commerce is in much attention nowadays, mainly due to its growing use. In today’s world, it has become imperative for companies to segment their customers and thereby take required measures to survive against other companies. Since there exist a lot, each company must fulfil the demands of their users or they might lose them to other alternatives that exist. This report aims to analyse the customer data from two years: 2018 and 2019 of the company: Autofurnish.com and thereby recommend methods to increase customer influx and give suggestions on what mistakes should not be repeated by the company in future for better performance and sales. To analyse the database, RapidMiner tool has been used. RapidMiner is an open source predictive analytic software that gives support regarding data mining. It lets the user to build models based on their needs and gives solutions quickly. For this analysis, K- means algorithm clustering will be used. Clustering is dividing groups based on similarities and K means is one of the very commonly used methods to do so. In the software, data has been imported having information about customers which is then analysed to prove different results draw a contrast between two years as told before keeping customer segmentation in mind based on various attributes given in the dataset. Relationships between the features are identified as assess to company’s performance.

Comparing Classification Algorithm With Genetic Algorithm In Public Transport Analysis

Congestion major cities in Indonesia caused by the proliferation of the use of private vehicles. Some people express their opinions and its opinion regarding public transport users through social media sites and other websites. This opinion can be used as a sentiment analysis material to find out whether the public transport service is positive or negative. The results of the sentiment analysis can help in the assessment and evaluation of the use of public transportation, it is also expected to improve services and facilities from public transportation so that the public tends to have a positive opinion. Based on the results of the sentiment analysis, it is expected that the community will switch to using public transportation which will certainly reduce congestion. In this study also added preprocessing stages by using the GataFramework framework to complete processes that cannot be done on RapidMiner tools. The method used in this study is sentiment analysis with the method of applying genetic algorithms for feature selection with comparative classification algorithms. Performed by testing the composition of various data. From the results of testing for the case in this study, it was found that the Support Vector Machine classification algorithm based on Genetic Algorithms had a fairly good average accuracy of 76.11% and AUC value of 0.778% with the Fair Classification diagnosis level compared to the three methods such as Naive Bayes, Support Vector Machine and Naive Bayes based on Genetic Algorithms. So that in this study Support Vector Machine classification algorithm based on Genetic Algorithm can be recommended as an algorithm classification good enough to analyze land transportation public sentiment. Based on the analysis it is expected that the public sentiment will switch to using public transport which would reduce congestion.

Comparation of K-Nearest Neighboor (K-NN) and Naive Bayes Algorithm for the Classification of the Poor in Recipients of Social Assistance

Poverty is a problems faced by developing countries, as well as Indonesia. According to data from the Central Statistics Agency in 2018, more than half the distribution of the poor population in Indonesia is in Java, which is 13,340.15 million people. Somokerto village is one of the villages in the district of Magelang, Central Java, which receives government assistance in an effort to reduce poverty. But in the process of classifying citizens who are entitled to receive assistance is still done manually. Manual classification is considered inaccurate in obtaining the results of social assistance recipients. In overcoming this problem, we need a systematic calculation to get accurate results. In this case, the researcher uses data mining classification calculation by comparing 2 calculation methods, namely K-NN and Naïve Bayes. The reseachers use Rapidminer tools. The research stages are identification of problems, data collection, implementation K-NN, Implementation Naive bayes, data testing process to produce accuracy and compare the result. The results obtained are the accuracy of Naïve Bayes higher than K-NN, namely Naïve Bayes 89.04% and K-NN 87.67%. This figure is classified in the category of good classification. From the results of the study it can be concluded the Naïve Bayes algorithm is suitable to be applied in the calculation of recipients of social assistance.

Comparative Analysis on Dimension Reduction Algorithm of Principal Component Analysis and Singular Value Decomposition for Clustering

Clustering is a method of dividing datasets into several groups that have similarity or the same characteristics. High-dimensional Datasets will influence the effectiveness of the grouping process. This study compares two dimension reduction algorithms, namely Principal Component Analysis (PCA) and Singular Value Decomposition (SVD) using K-Means clustering method to find out the best algorithm with the smallest Bouldin Davies Index evaluation. The dataset of this study involved public data from UCIMachine Learning which contains the number of weekly sales of a product. Data processing is performed by comparing the number of clusters from 3 to 10 and the dimension reduction from 2 to 10. From the data processing the RapidMiner tools, application with dimension reduction can provide better results than without dimension reduction. In particular, the PCA algorithm shows better results than the SVD, with which the best number of clusters is 5, and the number of dimensional reductions is 3 with a Bouldin Index of 0.376.

Analyzing Student’s Learning Interests in the Implementation of Blended Learning Using Data Mining

Blended Learning combines teaching and learning activities in the classroom and online teaching. In its implementation, this learning method requires a lot of data. One of them is the student's online test score data which can be used as an evaluation of learning. For this reason, in this study, data mining is used to determine the results of online student examinations as well as to determine student interest in learning about the implementation of Blended Learning. Data mining techniques are used in the logs of online learning session results, so that one can assess the online learning system used. By assessing the system, it can be identified which students who have studied hard and those who have not studied in the online exam. The series data used are student test score data on the State Polytechnic of Malang Learning Management System (LMS). The student score dataset is arranged based on variables in the Educational Process Mining (EPM) Dataset of UCI, which are obtained from teacher’s assignments. In addition, data mining classification is used to determine student interest in learning towards blended learning. In the process of data mining, comparative analysis is carried out using the features of the RapidMiner tool to carry out the process of student data for training and data validation. This process uses several algorithms along with student data which is divided into two sets for training and validation. From the results of data mining calculations produce a classification with minimum errors. From the test, the resulting linear regression algorithm has RMSE 0.000 and SE 0.000, while the neural network algorithm has RMSE 0.525 and SE 0.275.

Algoritma Apriori Sebagai Solusi Kontrol Persediaan Suku Cadang Mobil PT. Buanasakti Aneka Motor Jakarta

PT. BuanaSakti Aneka Motor is one of the car distributor agents located at warung buncit raya street No. 109 Duren Tiga, South Jakarta. Everyday the data piles is increase more and longer. PT. BuanaSakti Aneka Motor sometimes has difficulty to knowing how many spare parts occur in one transaction, so data or new stock information is difficult to find sometimes experiencing errors. To find out which products have the most sales and how they relate to one another, a priori algorithm method and the help of Rapidminer Tools are needed to produce associative rules. From the results of the discussion and analysis of the data it can be concluded that the application of a priori algorithm in determining the combination of itemset with a minimum support of 25% and a minimum of 60% confidence found 2 association rules, where the highest value of support and confidence is a radiator with a minimum stop lamp support of 26.92% while the minimum confidence is 64.65% and the oil filter with the minimum stop light support is 35.32% while the minimum confidence is 61.58%.

Customer Churn Prediction using Logistic Regression with Regularization and Optimization Technique

Customer Relationship Management (CRM) is a challenging issue in marketing to better understand the customers and maintaining long-term relationships with them to increase the profitability. It plays a vital role in customer centered marketing domain which provides a better service and satisfies the customer requirements based on their characteristics in consuming patterns and smoothes the relationship where various representatives communicate and collaborate. Customer Churn prediction is one of the area in CRM that explores the transaction and communication process and analyze the customer loyalty. Data mining ease this process with classification techniques to explore pattern from large datasets. It provides a good technical support to analyze large amounts of complex customer data. This research paper applies data mining classification technique to predict churn customers in three variant sectors Banking, E-commerce and Telecom. For Classification, enhanced logistic regression with regularization and optimization technique is applied. The work is implemented in Rapid miner tool and the performance of the prediction algorithm is assessed for three variant sectors with suitable evaluation metrics.

Penerapan Data Mining Dalam Pengelompokan Kota Berdasarkan Provinsi Yang Tanggap Terhadap Ancaman Narkotika Dengan Menggunakan K-Medoids

Narcotics is a problem that is not uncommon to be heard again by the public inside and outside the country. Especially at this time the development of Indonesia's population is very rapid so it is vulnerable to the threat of narcotics. There are several parties who take advantage of this development, especially the narcotics dealers, who are trying to destroy the younger generation by smuggling narcotics into the country and this is a potential market for illicit drug trafficking. The role of the community is very important in responding to this especially at the city / district level. To solve this problem the writer uses the K-Medoids method to classify cities that are responsive to the threat of narcotics. K-Medoids is one of the methods that exist in data mining by using clustering or grouping techniques. In this method the data that has been collected will be processed through the calculation process first by following the steps in the calculation process of the K-Medoids method that has been set so that it can get effective and accurate results. Using the K-Medoids method can find out the number of groupings in cities / districts that have high or low responsiveness. If a city has a low responsiveness to the threat of narcotics, it can help the City / District National Narcotics Agency (BNNK) to improve or add facilities such as UKS, socialize to the community and others in a City / Regency that has a low level of responsiveness to the threat narcotics, and increase public awareness of the dangers of narcotics. And if the city has a high response to the threat of narcotics, it is expected that the people in the city can maintain and increase their responsiveness (level of awareness) to the threat of narcotics. The results of grouping from the RapidMiner tools with the calsulation of k-medoids obtained a high clusters (C1) of 7 and a low cluster (C2)

Algoritma K-Means untuk Pengelompokkan Dokumen Akta Kelahiran pada Tiap Kecamatan di Kabupaten Simalungun

A birth certificate is a document that must be owned by a citizen. This document contains information about the birth of a person and is an official record of proof of state recognition of the person's existence. The purpose of this study is to group birth certificate documents in each district in Simalungun Regency. The research data was obtained from the Population and Civil Registry Office of Simalungun Regency. The grouping algorithm used is the K-means algorithm which is one of the Data Mining algorithms that is good for the case of grouping. By using this algorithm the data that has been obtained can be grouped into several clusters, where the application of the K-Means Clustering process uses the RapidMiner tool. Data is divided into 3 groups: high (C1), medium (C2) and low (C3). The results obtained from this study are in December entered into a high level cluster (C1), in August, September and October entered into a medium cluster (C2), and in January, February, March, April, May, June, July and November entered into the low cluster (C3).

IMPLEMENTASI DATA MINING UNTUK MENENTUKAN KELAYAKAN PEMBERIAN KREDIT DENGAN MENGGUNAKAN ALGORITMA K-NEAREST NEIGHBORS (K-NN)

The banking world in terms of lending to customers is routine activities that are at high risk. In its execution, the problematic credit or bad credit is often due to the lack of careful credit analysis in the process of granting credit, as well as from poor customers. The purpose of this study is to implement data mining to assist in conducting credit analysis process in order to produce the right information whether the customer who will apply for the credit is worthy or not to be able to see the potential payment by the customer. The attributes used in this study consist of 11 attributes i.e. marital status, number of liabilities, age, last education, occupation, monthly income, home ownership, warranties, loan amount, length of loan and description as a result attribute. The methods of data collection used are observation, interviews, and documentation. The method used in this study is K-Nearest Neighbor (K-NN). From the results of evaluation and validation using the K-5 fold that has been done using the RapidMiner tools obtained the highest accuracy results from the K-Nearest Neighbor (K-NN) method of 93.33% in the 5th test.

IMPLEMENTASI DATA MINING UNTUK MENENTUKAN KELAYAKAN PEMBERIAN KREDIT DENGAN MENGGUNAKAN ALGORITMA K-NEAREST NEIGHBORS (K-NN)

The banking world in terms of lending to customers is routine activities that are at high risk. In its execution, the problematic credit or bad credit is often due to the lack of careful credit analysis in the process of granting credit, as well as from poor customers. The purpose of this study is to implement data mining to assist in conducting credit analysis process in order to produce the right information whether the customer who will apply for the credit is worthy or not to be able to see the potential payment by the customer. The attributes used in this study consist of 11 attributes i.e. marital status, number of liabilities, age, last education, occupation, monthly income, home ownership, warranties, loan amount, length of loan and description as a result attribute. The methods of data collection used are observation, interviews, and documentation. The method used in this study is K-Nearest Neighbor (K-NN). From the results of evaluation and validation using the K-5 fold that has been done using the RapidMiner tools obtained the highest accuracy results from the K-Nearest Neighbor (K-NN) method of 93.33% in the 5th test.

Twitter text mining for sentiment analysis on government’s response to forest fires with vader lexicon polarity detection and k-nearest neighbor algorithm

Opinions about the government’s response to forest fires have drawn many opinions from the community. One way for people to express their opinions is to use social media Twitter. This study conducted a sentiment analysis process on the government’s response to handling forest fires in Indonesia in 2019 with data sources from Twitter. The analysis was carried out on 6325 datasets written on Twitter on September 20, 2019, and then through the process of pre-processing, automating labeling and classification. The automate labeling process uses a Vader that automatically detects the negative or positive polarity of each data and then goes through the classification process using the KNN algorithm. The test results that were built using rapidminer tools showed an accuracy level of the KNN algorithm of 79.45%, the highest if compared to other classifier algorithms such as decision trees, naïve Bayes and random forests. The sentiment analysis process can almost run automatically without human touch because there is already automated labeling using Vader. Testing sentiment analysis related to the government’s response to forest fires can be analyzed using the KNN algorithm and lexicon polarity detection Vader can be done properly.

Pemanfaatan Metode Association Rules dan Holt-Winter Multiplicative untuk Meningkatkan Peluang Penjualan Obat Pertanian

The competition level between companies on executing product marketing is rapidly increasing, so the companies have to understand the importance of correlation between external environments of company with consumer’s needs. One of the efforts that can be done is by utilizing data warehouse and the application of infrastructure in information and technology field. This research combined Association Rules method to extracting pattern and finding every possibility that potential to increase sales and Holt-Winter Multiplicative method to estimate the alteration of trend on the seasonal data. After passed through data processing process by using RapidMiner tools, information that consists of correlation pattern between rule that describe the comparison of product and the sales working area and season that affects the product sale. The pattern used by company to know which product is often purchased by customer. Besides that, this research produces changing trend data of PT ABC’s product that generated by result of previous data comparison with forecast data. Based on value of error rate Mean Absolute Percentage Error (MAPE) in estimating forecast result on the PT ABC’s sales transaction data during 3 years, it shows good level of accuracy. Result of data test, by considering rule that formed and forecast result so the company can control and manage product in order to avoid incorrect sales. This thing will effect on repression of operational cost and PT ABC can identify available opportunities to increase sale of agricultural medicine.

Application of Data Mining Technique using K-Medoids in the case of Export of Crude Petroleum Materials to the Destination Country

The purpose of this research is to analyze and implement the data mining technique in the export of crude petroleum materials to the destination country. This is because Indonesia is a member of OPEC (Organization of the Petroleum Exporting Countries) which is one of the largest petroleum exporters in the world. It aims to obtain profits in the form of foreign exchange income obtained by the State. The data source was obtained from the Central Statistics Agency (BPS) with the website https://www.bps.go.id for data for 2017-2018. The calculation process The technique used is clustering with the K-Medoids algorithm. The calculation process is carried out using the help of Rapid Miner tools. The clusters used in this study are 2 namely: high cluster (C1) for export of crude oil materials and low cluster (C2) for crude oil materials. The results of the study stated that the high cluster (C1) consisted of 3 countries (Japan, Thailand and the United States) and the low cluster (C2) consisted of 6 countries (South Korea, Taiwan, China, Singapore, Malaysia and Singapore). It is hoped that the research results will be input and information for the government to rearrange policies in order to increase competitiveness, ensure business certainty and the sustainability of domestic industrial raw materials.

Penerapan Aturan Asosiasi Algoritma Apriori Terhadap Penjualan Thai Tea Pada PT Nyonya Besar Lestari

Competition in the business world, especially in the increasingly lively beverage sales, requires the sellers to find strategies to increase sales. More and more sales data every day can be used to develop marketing strategies if processed correctly. To find out the most sales Apriori Algorithm is needed to be able to find out and with the help of Rapidminer tools, products with the most sales can be known. Apriori algorithm including the type of association that has raised the attention of many researchers to produce an efficient algorithm is the analysis of high frecuency patterns (frequent pattern mining) in this study can be implemented on the interest of buying drinks with the data used are 31 transactions with a minimum support or 30% and minimum confidence 60%.

KOMPARASI ALGORITMA DECISION TREE, NAIVE BAYES DAN K-NEAREST NEIGHBOR UNTUK MEMPREDIKSI MAHASISWA LULUS TEPAT WAKTU

Private Universities (PTS) compete so tight in providing performance in producing quality graduates. In addition, the number of universities in Indonesia which counts a lot both PTN and PTS makes the higher competition between universities as well. So the university strives to improve quality and provide the best education for service recipients, namely students, where one of the problems if there are some students who are late graduating or not on time so that it becomes an obstacle to the progress of the college. Prediction of students graduating on time is needed by university management in determining preventive policies related to early prevention of Drop Out (DO) cases. This prediction aims to determine the academic factors that influence the period of study and build the best prediction model with Data Mining techniques. There are 11 attributes used for Data Mining Classification, namely NPM, Gender, Age, Department, Class, Occupation, Semester 1 Achievement Index, Semester 2 Achievement Index, Semester 3 Achievement Index, Semester 4 Achievement Index and Information as result attributes. From the results of evaluations and validations that have been carried out using the RapidMiner tools the accuracy of the Decision Tree (C4.5) method is 98.04% in the 3rd test. The accuracy of the Naïve Bayes Method is 96.00% in the 4th test. And the accuracy of the K-Nearest Neighbor Method (K-NN) of 90.00% in the second test.

K-means Clustering Analysis of Crimes on Indian Women

Violence against women is seen as sexual or physical activity committed against women. In India, general forms of violence against women in India includes cruelty by relatives, dowry, rape, sexual assault, kidnapping, immoral trafficking, molestation etc. The security of the women is the utmost priority of any government in this world. In India, many policies and laws have been enforced to ensure the safety against women. Technology is being the biggest supporter to the government in this context. Data mining allows various techniques such as clustering classification, regression provides analysis in any form of data and helps intelligent predictions on the given dataset. In this paper, we use k-means clustering analysis on women crime dataset. As a part of pre-processing, we collated the data entries which had crime cases against women and made women crime sub-dataset from the real dataset. We then applied K means clustering for further analysis. We used a rapid miner tool for clustering analysis as it is widely used for clustering purposes. After completion of clustering analysis, we proposed our views and discussions on the clustering results. At the end, we ended up giving the futuristic work to be further done on the derived dataset we made and made available on public repositories.

Application of Feature Weighting for the Intensification of Data Classification

Classification is the supervised learning technique which is applied in many of the real time applications. In this study we have considered three classifiers which are widely used and then the intensification of the classifiers are considered. Among various methods to improve the performance of the classifiers, this research concentrate on the feature weighting techniques applied for the classifiers. This analysis is done based on the results obtained from the Rapidminer tool. Here we have deployed four feature weighting techniques for the intensification of the three classifiers. It is tested with three dataset. The experimental environment and the results are discussed in detail.

Prediksi Kelulusan Mahasiswa menggunakan Algoritma Naive Bayes (Studi Kasus 5 PTS di Banda Aceh)

The e-questionnaire application that researchers built using CodeIgniter and React-Js This study aims to data mining by using rapidminer tools to collect student data from the Feeder application page from the class of 2010-2014 which is assumed that the student class has been declared graduated in 2018. The data was collected from 5 (five) Private Universities in the City Banda Aceh. then by observing the graduation level using data mining can bring a considerable contribution to educational institutions, in an effort to improve curriculum competency in Higher Education, it is expected that the results of data mining can make reference to curriculum standards as a form of graduate competency improvement. The research method uses the Cross-Industry Standard Process for Data Mining (CRISP-DM) which is used as a standard data mining process as well as a research method with stages starting from Business understanding, data understanding, data preparation, modeling, evaluation, and deployment. The results showed that the data mining algorithm for graduation prediction based on the selected pass accuracy attribute revealed that the prediction level was uniform with the algorithm used, Naïve Bayes, prediction accuracy was 84%. The data attributes that were found to have significantly influenced the classification process were the GPA and Study Length. The results obtained that students who graduated by 60% are students who are educated in ASM Nusantara and AMIK Indonesia, while in Banda Aceh STIES and Serambi University Mecca the prediction of graduation is 52%. Another thing is different from STIA Iskandar Thani where the prediction of graduating is only 48% and not passing on time is 52%. The results of this prediction can reveal and become a recommendation for prospective students or academics to increase the quantity of graduates and increase student confidence in tertiary institutions.Keywords:Prediction, Student Graduation, Naive Bayes Algorithm.

Classification of Operational and Financial Variables Affecting the Bullwhip Effect in Indian Sectors: A Machine Learning Approach

Background: The increased variability in production or procurement with respect to less increase of variability in demand or sales is considered as bullwhip effect. Bullwhip effect is considered as an encumbrance in optimization of supply chain as it causes inadequacy in the supply chain. Various operations and supply chain management consultants, managers and researchers are doing a rigorous study to find the causes behind the dynamic nature of the supply chain management and have listed shorter product life cycle, change in technology, change in consumer preference and era of globalization, to name a few. Most of the literature that explored bullwhip effect is found to be based on simulations and mathematical models. Exploring bullwhip effect using machine learning is the novel approach of the present study. Methods: Present study explores the operational and financial variables affecting the bullwhip effect on the basis of secondary data. Data mining and machine learning techniques are used to explore the variables affecting bullwhip effect in Indian sectors. Rapid Miner tool has been used for data mining and 10-fold cross validation has been performed. Weka Alternating Decision Tree (w-ADT) has been built for decision makers to mitigate bullwhip effect after the classification. Results: Out of the 19 selected variables affecting bullwhip effect 7 variables have been selected which have highest accuracy level with minimum deviation. Conclusion: Classification technique using machine learning provides an effective tool and techniques to explore bullwhip effect in supply chain management.