| Summary: | One of the government's land management based on article 19 of Law - Basic
Agrarian Law is the provision of legal certainty to the subject of rights through land
registration activities. Acceleration of land certification activities conducted by government
for all land parcels in the territory of the Republic of Indonesia can be registered. Efforts
made to complete certification in Indonesia is increasing land services based informatics /
Land Office Computerization (LOC), shuttle service Larasita ball (People's Service for
Certification of Land) and the strengthening of rights - the right of the people of the land
mass through self-certification. This research aims to design an evaluation system of mass
self-certification activities at the District Land Office Purworejo and know the level of success
of the activities in question.
Evaluation system used in this research is artificial neural network backpropagation
training. The basic principle of backpropagation training is to make the learning of nodes
included in nerve tissue, so as to achieve the smallest error and most close to the target set.
Nodes are included in the neural network is a set of data which is a combination of the
weights of each variable. Learning is done in two phases: forward propagation and back
propagation stage. Calibration on the research data sets were selected after the learning
process is over, the testing process is in principle is a step forward propagation using
weights and biases values last produced at the time of learning.
The results of this research, the activity of mass self-certification in the District
Land Office Purworejo categorized the level of success in the medium. Methods of training
backpropagation neural networks are able to evaluate the mass self-certification activities
and produce a relatively small error values against the target set.
|
|---|