Download Bukti Potong
| Nama |
|---|
| NIP |
|---|
| NPWP |
|---|
| Alamat |
|---|
| Tahun |
|---|
The turnover is a matter that is not desired by the company, this is a classic problem faced by entrepreneurs from the industrial revolution. Poor working conditions, wages are too low, hours of work over the limit and the absence of jaminansosial is the main cause of turn over. Turn over the place cost the company both in terms of cost, resources, and motivation of employees, the company lost a significant amount of labor that must be replaced with new employees. Companies ranging from recruitment to pay up to get ready workforce. During the vacancy then existing labor sometimes do not correspond to the task at hand so that it becomes dormant. Remaining employees will be affected the motivation and morale. Employees who previously did not try to find a new job will start looking for a job, which would then make the turn over. This situation clearly needs to take a loss because it sought to solve. PT. Rig Tenders Indonesia Jl. Belitung no. 88 Banjarmasin, since October 2009 until December 2011, of the 531 people still working there were about 293 people stopped working, during the period on average get 13 people who stopped working a month.
Bayesian theory is a theory that takes into account the condition probability of probability of an event (hypothesis) depends on another event (evidence). Basically, the theorem says that future events can be predicted on the condition precedent has occurred. Application of the Naive Bayes algorithm in determining the status of employee turnover at PT. Tender Rigs Indonesia conducted by grouping the data based on specific attributes, the study uses namapg column, CPA, ota, statp, JTG, jabata, umura, lamaker and statp, where the columns of BPA, ota, jabata, umura, and lamaker done so as to produce the classification better accuracy. Test results using 824 data by adding the PSO optimization Naive Bayes algorithm generates higher accuracy of 2.65% compared with no use PSO optimization is only a level of accuracy of 73.43%, while from the test results using 824 data by adding the PSO optimization Naive Bayes algorithm produces an accuracy of 76.08%, was higher accuracy value of 5.48% compared to manual calculations without using PSO optimization by using the 807 training data and test data 17 is only a level of accuracy of 70.6%. The final results of the ROC (Receiver Operating Characteristic) Curve obtained AUC values (Area Under the Curve) of 0.8, including a very good classification.
Keywords: Classification, status turn over, Naive Bayes algorithm, Particle Swarm Optimization
