الفهرس 
IntroductionOnly 14 pages are availabe for public view
 |  | from | 128 |  |  |
| | | from | 128 | | |
Abstract
Decision Support Systems (DSS) is a particular type of computerized information system that supports business and organizational decision making activities. On the other hand, Data Mining (DM) supports a decision through discovering knowledge from raw data which is hidden. Data mining techniques is considered the best and accurate technique for building a DSS system concerned to loans in some loan systems in one of the Arab countries. Loans are provided to people for such critical circumstances. There are various types of loans like home loans, personal loans, student loans and so on. Lenders can be private or governmental authority as Banks. Lenders need a strong system to support their decisions by lending people and get the most benefit with less money losses. This thesis aims to achieve a set of objectives, including the design of an accurate system that supports the decision on borrowers’ loans as well as knowledge of the obstacles faced by borrowers in returning the amounts of loans they have received. These objectives can be achieved by building a comprehensive and balanced data set through following the following points. Loan data were collected from six regions covering the country, • Data was collected as historical data for ten (10) years ago (1420/1999 -1430/2008). • Total number of contracts were determined in every year of those ten ، years for each region • The percentage of contracts in every region was determined according to all contracts issued on all the regions of that year. • ETL (Extract, Transform and Load) operations are made for number of records according to the region percentage Then, a mining technique is constructed on that data set to predict the borrowers state as Safe, very safe, risk , very risk , Hence, a decision maker can take the right decision according to that result. The thesis present high accurate rich to 85.4% with Naïve Bayes algorithm and 89% with Decision tree algorithm.