Patthama Saeting.. A comparative study of book selling recommendation techniques. Master's Degree(Technology of Information System Management). Mahidol University. Mahidol University Library and Knowledge Center. : Mahidol University, 2013.
A comparative study of book selling recommendation techniques
Abstract:
In todays highly competitive world, quick decisions should be made without major mistakes. The decisions based solely on individuals are frequently complicated and are also more time-consuming in order to obtain the optimal result. These possibly cause errors as a result of several negative effects on businesses. Product distribution is a key problem in logistic system, such as how to manage inventory to satisfy customers need without affecting product quality. Therefore, it is essential to have a plan for efficient product distribution, and how to control the quantity of the product. Product recommendation is another important problem when introducing a new products line that has arrived. Manufactures and retailers do not know which customers would be interested in the new products; but if we distribute product evenly to every customer, and wait for responses, that will waste time in the distribution of the products. This research proposes three methods that have standard or are used in Recommender System; Collaborative Filtering, Artificial Neural Network, and Decision Tree., for solving problems of book distribution and book recommendations for customers who would like to be bookstore owners by using a case study of book distribution from Kledthai Book Distribution Company. Input data was divided into 2 different types which were comprised of numeric data and nominal data. The result will be displayed in the form of prediction that will help monitor the result and determine errors in predicting quantity of sales. According to the experiment, a decision tree method was shown to be the most accurate of both data types, which has an accuracy at 40.10% of input data that is numeric data and 40.58% of input data that is nominal data. This researcher concludes that the decision tree method has the highest accuracy value, or was able to yield the most accurate prediction
Mahidol University. Mahidol University Library and Knowledge Center