International Journal of Academic Engineering Research (IJAER) ISSN: 2000-001X Vol. 3 Issue 2, February – 2019, Pages: 21-25 www.ijeais.org/ijaer 21 A Proposed Artificial Neural Network for Predicting Movies Rates Category Ibrahim M. Nasser, Mohammed O. Al-Shawwa, Samy S. Abu-Naser Department of Information Technology, Faculty of Engineering and Information Technology Al-Azhar University- Gaza, Palestine Azhar.ibrahimn@gmail.com mohammed.o.alshawwa@gmail.com Abstract: We proposed an Artificial Neural Network (ANN) in this paper for predicting the rate category of movies. A dataset used obtained from UCI repository created for research purposes. Our ANN prediction model was developed and validated; validation results showed that the ANN model is able to 92.19% accurately predict the category of movies’ rate. Keywords: Data Mining, Classification, Predictive Analysis, Artificial Neural Networks, movies classification 1. INTRODUCTION Artificial neural networks (ANNs) are, similar to our neural networks and offer a relatively good technique, which solves the problem of classification and prediction. ANN is a collection of mathematical models, which can simulate characteristics of biological neural systems and have likenesses with adaptive human learning. ANNs made of connecting processing elements called neurons, connected by links, which contain weight coefficients that are, playing the role of synapses in our neural system. The neurons often come in three layers: input layer, one or more hidden layers and output layer, (ANN Architecture is shown in figure (1)). ANNs handle data as biological neural networks, in addition, ANN has the possibility of recalling, learning and eliminating errors, and high speed of getting the solution, [1] because of that, the neural networks can be used for solving complex problems, like classification and prediction [2]. ANNs were effectively applied in variety of applications for solving difficult and real problems [3]. ANN were found to be more efficient and more accurate than other classification techniques [4]. Classification by a neural network is done in two separate phases. First, the network is trained on a dataset. Then the weights of the connections between neurons are fixed so the network is validated to determine the classifications of a new dataset [5] . In this paper, we used about 70% of the total sample data for network training, and 30% for network validation. While many models of ANNs have been proposed, the feed- forward neural networks (FNNs) are the most common and broadly used in many applications. Mathematically, the problem of training an FNN is the minimization of an error function E ; In another word, to find a minimizer w = (w 1 ,w 2 ,…, w n ) such that w = min E(w), where E is the batch error computed by the sum of square differences over all examples of the training dataset.   ∑ ∑(      )          is the output of the j-th neuron that belongs to the L-th (output) layer, N L is the number of neurons of the output layer, t j,p is the anticipated response at the j-th neuron of the output layer at the input pattern p , and p represents the total number of patterns which used in the training dataset. A traditional technique to solve this problem is by an iterative gradient-based training process, which produces a series of weights {  } starting from an initial point      Using the iterative formula          where k is the current iteration,    is the learning rate and   is the decent search direction [5]. Our study main purpose is to develop a neural network as classification technique to predict the category of movies rate. A dataset from UCI repository [6,7] was used for this purpose. Figure 1: ANN Architecture