International Journal of Computer Techniques

08 - 14

Authors: Saruni Dwiasnati, Yudo Devianto

In order to deal with business competition and increase the company's revenue, both company leaders and management in a company are required to be able to make the right decisions in determining the sales strategy. To be able to do this, companies need a large amount of information to be analyzed further. The company executive expects technology that can produce information that is ready to be used to assist them in making corporate strategic decisions. They want to know what products should be improved, how much the results obtained by the company will be achieved. To meet the needs of entrepreneurs above, there are many ways that can be pursued. One of them is by utilizing corporate data. Based on the identification of the problems that have been described, the problem is formulated as follows: how the optimization results for the Naive Bayes algorithm based on Particle Swarm Optimization, how to build a prototype to predict customer data by using the appropriate data mining classification method. This study aims to recommend a prototype of the data mining classification method for potential customer predictive data in offering an existing product based on data mining classification method, namely the Na�ve Bayes algorithm based on Particle Swarm Optimization. This study, if viewed from the form of data and information that is managed, this research is classified as a quantitative type of research. Quantitative research is a research whose hypothesis can be tested by statistical techniques. This method is used when conducting quality testing using the k-fold cross validation method (k = 10) which displays the value of accuracy, precision, recall, ROC for each method that is compared. The results of application development and model performance measurement will be explained in this section. Application development will be discussed in testing to show that the results of the application are made as expected. Whereas in the measurement of model performance will be explained the results of the model performance measurement for the analysis of potential customers and potential customers not to be offered insurance products at PT. XYZ by using Naive Bayes based on Particle Swarm Optimization, the results obtained in the algorithm that has been optimized are made an application developed.

21 - 27

Authors: Kalai Mani.A, Girija Magesh

Gardens make up the biggest inundated yield by surface zone in North America, and convey with it an interest for more than 9 billion gallons of freshwater every day. In spite of ongoing improvements in water system control and sprinkler innovation, cutting edge water system frameworks do nothing to make up for regions of turf with heterogeneous water needs. In this work, we overcome the physical constraints of the customary water system framework with the improvement of a sprinkler hub that can detect the neighbourhood soil dampness, impart remotely, and incite its own sprinkler in view of a midway processed calendar. A model is then created to process dampness development from overflow, assimilation, and dispersion. Incorporated with an improvement system, ideal valve booking can be found for every hub in the space. In a turf territory covering more than 10,000ft, two separate arrangements traversing a sum of 7 weeks demonstrate that MAGIC can diminish water utilization by 23.4% over customary grounds booking, and by 12.3% over condition of-the art evapo-transpiration frameworks, while significantly improving conditions for plant wellbeing. Notwithstanding ecological, social, and medical advantages, MAGIC is appeared to restore its speculation in16-year and a half in light of water utilization alone.

34-41

Authors: Reza Sulistyawan, Mohammad Syafrullah, HadidtyoWisnu Wardani, Novian Hendrianto

Employee performance indicates the result of work from employee in conducting the job based on competence, manner, and motivation. In evaluating employee performance, there are many factors, including loyalty, responsibility, discipline, integrity, team work, and leadership factors. Employee roles in advancing the organization is needed because without a good performance, the organization cannot achieve its objectives well. In AMIK PakartiLuhurTangerang, employee performance evaluationis done subjectively from the result of meeting, hence, the result of evaluation is not accurate. The samples is done by purposive sampling technique, which is the questionnaire is given to respondents. Employee Performance Evaluation Model is made with Adaptive Neuro Fuzzy Inference System (ANFIS) Sugeno method and Toolbox Matlab 8.2 R2013b.To ensure that the system meet the minimum standard quality, thus, the system use Software Quality Assurance (SQA) in evaluating software quality.The conclusion from employee performance evaluation model has a good level of accuracy.

48-53

Authors: Nur Ani, Handrie Noprisson

Activities of electronic business in the online travel booking application are ultimately aimed to increase revenue as much as possible through an increase in the number of purchases. To obtain more revenue is influenced by many factors, both from the consumer perspective and from the business organization itself. This study is a preliminary study to develop a factor model that affects purchase intention in online travel booking application. The research model examined four main variables, i.e. innovativeness, trust, perceived value and purchase intention. This model is adapted from research model that is proposed by Escobar-Rodriguez (2017). We used bootstrap resampling procedure based on partial least squares (PLS) by processing 984 data.Based result of data analysis, trust has the biggest influence to purchase intentionwith t-value 8.280. Moreover, trust also support perceived value with t-value 29.619. Perceived value influences the purchase intention in online travel booking application with t-value 7.091.

60-66

Authors: Badrus Zaman Sapto Mardi Putro, Krisna Adiyarta M, Sri Wahyuningsih, Pepi Permatasari

The selection process for PPA Scholarship students currently at Pakarti Luhur AMIK is not subjective, namely through meeting results so that the accuracy of the results is reduced. This decision is adjusted by several requirements criteria so that the selection of PPA scholarship recipients in accordance with the requirements contained in the Pakarti Luhur scholarship can be chosen, which is based on the Academic Achievement Scholarship (PPA) guidelines, circulated by the Ministry of Research, Technology and Higher Education. This study is in the form of a case study with research and development methods whose subject of research is the Information and Computer Management Academy of Pakarti Luhur. This model is designed for the selection process of students who have the opportunity to win PPA scholarships, with the Naive Bayes method used. The results of the model with the Naive Bayes Method in the process of selecting desktop-based PPA Scholarship winners display features of data class input menu, input data requirements, student input to be selected, data conversion process to training data, validation process, and data testing. The results of the model design can provide objective results and ease in helping to carry out the selection process. The test results also show that the use of the Naive Bayes method can be used as an alternative in making decisions in the selection of PPA Scholarship students who have an accuracy of 93.4783%.

74-78

Authors: UmniySalamah , DesiRamayanti

Social media is usually used to express emotion in opinion or complaint statement regarding an object. In organization, this statement contained opinion and complaint to a service or program by it organization. This feedback can be used by organization to improve and enhance their quality. This research attempted to classify the reports from LAPOR! based on complaint and non-complaint using machine learning algorithm named Logistic regression (LR). This research completed through five research phases. The first phase is data collection, which is followed by crawling and labelling dataset, procession data, modelling dataset using Logistic Regression (LR), and evaluating classifier performance. As the result, the Logistic Regression model using Count Vectorizer feature extraction achieved better performance than Tfidf Vectorizer.

85 - 90

Authors: Mr. Umesh Nanavare , Prof.Darshana Patil

Textual information is mainly classified into two main types: facts and opinions. The entities, events and their properties are objective expression of facts. Facts are something which already happened or actually the case itself. They are (e.g., iPhone is an Apple product).Opinions are subjective expressions which describe peoples? judgment, viewpoint, feelings towards entities, events and their properties. (e.g., I like this Apple iPhone 6). Opinion Mining is the computational study of people�s opinions towards entities and their aspects. Entities basically refer to individuals, events, topics, products and organizations. Aspects are attributes or components of entities. In the last few years, social media is the main source to express and share people�s opinion on entities and their aspects. Because of the availability of vast opinionated web contents in the form of comments, reviews, blogs, tweets, status updates, etc. It is difficult for people to analyse all opinions at a time to make good decisions. There is a need for effective automated systems to evaluate opinions and generate accurate results. Mining opinion targets and opinion words from online reviews is important tasks for fine-grained opinion mining, the key component of which involves detecting opinion relations among words. To this end, this work proposes a Complementary Aspect based Opinion Targets with Opinion Words Co-Extraction Algorithm. Experimental Result shows the proposed algorithm extracts opinion targets and opinion words more precisely.