Abstract— This papers presents bayes network based classification method to evaluate the reusable software components and in identification of reusable components from existing legacy systems. The bayes network classifier algorithm takes a database and an attributes ordering as input and constructs a belief network structure as output. So many discrepancies exist between expert opinion and empirical data reported in Morisio et.al.’s recent TSE article. But the result of this evaluation depends on the probability of different instances. We find some difference related to factors that makes the success of software reuse. This implementation describes how those differences are detected and how the instances give true positive value and accuracy. Keywords—Bayes Network, Software reuse, TP, Accuracy. I. INTRODUCTION LASSIFICATION is a basic method in software reusability and data analysis that requires the construction of classifiers, i.e. a function that assigns a class to instances described by a set of attributes. In the recent years the researchers of software reusability gives Success and Failure Factors in Software Reuse sought key factors that predicted for successful software reuse. Their data comes from the high level management of a 24 department of a company. Bayes networks (BNs), also known as belief networks belong to the family of probabilistic graphical models (GMs). These graphical structures are used to represent knowledge about an uncertain domain. In particular, each node in the graph represents a random variable, while the edges between the nodes represent probabilistic dependencies among the corresponding random variables. These conditional dependencies in the graph are often estimated by using known statistical and computational methods. Hence, BNs combine principles from graph theory, probability theory, computer science, and statistics [6,7]. A Bayesian network   , , A N B is a directed acyclic graph (DAG) A N , where each node N n  represents a domain variable (eg, a Amanpreet Singh is student of Guru Nanak Dev Engineering College, Ludhiana. Punjab (India).( e-mail: dhanoa.a@gmail.com) Prof. Amanpreet Singh Brar is Head of Department of Guru Nanak Dev Engineering College, Ludhiana Punjab (India) (e-mail: appi_brar@yahoo.co.in). Dr. Parvinder Singh Sandhu is Professor, Rayat-Bahra Group of Institutes, Punjab (India) (e-mail: parvinder.sandhu@gmail.com). dataset attribute), and each arc A a  between nodes represents a probabilistic dependency, quantified using a conditional probability distribution   i  for each node i n . A BN can be used to compute the conditional probability of one node, given values assigned to the other nodes; hence, a BN can be used as a classifier that gives the posterior probability distribution of the class node given the values of other attributes. A major advantage of BNs over many other types of predictive models, such as neural networks, is that the Bayesian network structure represents the inter-relationships among the dataset attributes. Human experts can easily understand the network structures and if necessary modify them to obtain better predictive models. By adding decision nodes and utility nodes, BN models can also be extended to decision networks for decision analysis [11]. In case of the two-cluster based problem, the confusion matrix has four categories: True positives (TP) are Projects correctly classified as Successful cases of software Reuse. False positives (FP) refer to Unsuccessful or Failure projects incorrectly labeled as Success. True negatives (TN) correspond to Unsuccessful or Failure projects correctly classified as such. Finally, false negatives (FN) refer to Successful Projects incorrectly classified as failure as shown in table 4.1. TABLE I CONFUSION MATRIX OF PREDICTION OUTCOMES. Real Data Value of Project Status Predicted Project Success failure Success TP FP failure FN TN With help of the confusion matrix values the precision and recall values are calculated described below: A. Precision The Precision is the proportion of the examples which truly have class x among all those which were classified as class x. The technique having maximum value of probability of detection and lower value of probability of false alarms is chosen as the best prediction technique. A Bayes Network Based Classification Approach for Evaluation of Success of Software Reuse Amanpreet Singh, Prof. Amanpreet Singh Brar, Dr. Parvinder Singh Sandhu C International Conference on Machine Learning and Computer Science (IMLCS'2012) August 11-12, 2012 Phuket (Thailand) 175