ANN : Model of BackPropagation Architecture on the Logs Production by Type of Wood

Indonesia is rich in forest products. The production forest is a forest area that can be utilized for the community, such as logs, rattan, and some forest plants that have high economic value. This research aims to make the best architectural model by using artificial neural network. The method used is backpropagation algorithm. With this model it will continue to predict the output of logs. Data are sourced from BPS-Statistics Indonesia. Based on the results of research results of logs production using backpropogation method, obtained the result of 3 model architecture (18-18-1, 18-25-1 and 18-18-251) that model of architecture 1825-1 is the best model with 72% accuracy, MSE: 0.0221670942 and epochs: 660.


Introduction
Indonesia is rich in forest products, but not all forests can be harnessed and produced.For forest utilization, the government has established certain forest areas designated as Production Forest.The production forest is a forest area whose results can be utilized for the community, such as logs, rattan, and some forest plants that have high economic value.Logs that are cut or harvested can be used as raw materials for upstream wood processing production.This production of logs is produced from natural forests through the activities of forest concession companies.Forest Concessionaire Enterprises is a business / legal entity engaged in the field of harvesting of forest products.Based on the source of Indonesia's Forestry Statistics Book, the data of log production in Indonesia for the last 10 years has increased by 314.64%In 2001 Indonesia's log production amounted to 11,432,501 m3 and in 2011 amounted to 47,429,335 m3.Increased production of logs gives some negative impacts.One of them is the destruction of natural ecosystems that encourage the emergence of concerns shortage of industrial raw materials in the future.This study aims to create a prediction model of the total prediction of logs in Indonesia using artificial intelligence.This model can then be used to predict long-term log production so that the government can anticipate and provide solutions to the negative impacts resulting from the production of logs in Indonesia.One branch of computer science related to prediction is artificial intelligence.There is a lot of artificial intelligence that deals with predictions.The Artificial Neural Network is one of the branches of artificial intelligence [1].Today, the AI is a very important discipline and it includes a number of well-recognized and mature areas including Expert Systems [2][3][4], Fuzzy Logic [5][6][7][8], Genetic Algorithms [9][10][11], Language Processing, Logic Programming, Planning and Scheduling, Neural Networks and Robotics [12].There are many techniques that can be used for the implementation of the Neural Network Tirua one of the methods used is Backpropagation [2][3][4][5][6].Backpropogation is one of the artificial neural network algorithms that is often used to solve complex problems related to input identification, prediction, pattern recognition, and so on.Repeated training will result in a network that responds correctly to all its inputs.

Artificial Intelegence
AI is a field of study based on the premise that intelligent thought can be regarded as aform of computation -one that can be formalized and ultimately mechanized.To achievethis, however, two major issues need to be addressed.The first issue is knowledge representation, and the second is knowledge manipulation [1].

Artificial Neural Networks (NN)
Artificial Neural Network (ANN) is a computational model, which is based on Biological Neural Network.Artificial Neural Network is often called as Neural Network(NN) (See Figure 1).From Figure 1, to build artificial neural network, artificial neurons, also called as nodes, are interconnected [13,14].An artificial neuron is an abstraction of biological neurons and the basic unitin an ANN [15,16].The Artificial Neuron receives one or more inputs and sums them to produce an output [17,18].After successful training, user can give unlabeled data to be classified.

Architecture of Backpropogation
The back-propagation learning algorithm (BPLA) has become famous learning algorithms among ANNs.In the learning process, to reduce the in accuracy of ANNs, BPLAs use the gradient decent search method to adjust the connection weights.The structure of a backpropagation ANN is shown in Figure 2 [19].Each of these layers must be either of the following: 1. Input Layer -This layer holds the input for thenetwork 2. Output Layer -This layer holds the output data, usually an identifier for the input.3. Hidden Layer -This layer comes between the inputlayer and the output layer [20].Source: BPS-Statistic Indonesia

Output Data
The expected result is the selection of an arsitertur model to predict the results of the Logs Production by Type of Wood.The best architectural model is seen from the smallest minimum error rate.In this study, the minimum error used is: (0,001-0,009 or (-0,001)-(-0,009): True) and (>0,009: False).

Results
By using the same parameters on each activation function, namely: sigmoid bipolar (tansig) with net.trainparam.epochs= 1500000 net.trainparam.LR = 0.1; net.trainParam.goal= 0.001; net.trainParam.show= 1000; Minimum error of 0.001-0.009obtained result backpropagation architecture model 18-25-1 is the best with the accuracy of 72%.Here is the comparison of 3 architectural models and the performance of each architecture model

Conclution
Based on the results of research in determining the architecture model on the prediction of the result of logs production using backpropogation method, obtained the result of 3 model architecture (18-18-1, 18-25-1 and 18-18-25-1) that model of architecture 18-25-1 is the best model with 72% accuracy.

Figure 1 .
Figure 1.Logs Production in Indonesia (Source : Indonesia Forest Statistics Book)

Table 1 . Logs Production by Type of Wood (M3)
The data used in this researc-2015)ound wood production data by type obtained fromBPS Statistic Indonesia (2004-2015).Log production data can be seen in the following table: 3.1.ResearchFrameworkA framework of research work used in solving this research problem.