• Journal of the Korea Society of Computer and Information
• /
• v.10 no.1 s.33
• /
• pp.45-52
• /
• 2005

We proposed the product recommendation algorithm mixed the temporal association rule and the exponential smoothing method. The temporal association rule added a temporal concept in a commercial association rule In this paper. we proposed a exponential smoothing temporal association rule that is giving higher weights to recent data than past data. Through simulation and case study in temporal data sets, we confirmed that it is more Precise than existing temporal association rules but consumes running time.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.31B no.8
• /
• pp.211-218
• /
• 1994

In this paper we develope a hybrid learning rule to improve the robustness of multi-layer Perceptions. In most neural networks the activation of a neuron is deternined by a nonlinear transformation of the weighted sum of inputs to the neurons. Investigating the behaviour of activations of hidden layer neurons a new learning algorithm is developed for improved robustness for multi-layer Perceptrons. Unlike other methods which reduce the network complexity by putting restrictions on synaptic weights our method based on error-backpropagation increases the complexity of the underlying proplem by imposing it saturation requirement on hidden layer neurons. We also found that the additional gradient-descent term for the requirement corresponds to the Hebbian rule and our algorithm incorporates the Hebbian learning rule into the error back-propagation rule. Computer simulation demonstrates fast learning convergence as well as improved robustness for classification and hetero-association of patterns.

• The KIPS Transactions:PartD
• /
• v.18D no.6
• /
• pp.423-432
• /
• 2011

In this paper, we introduce XOnto-Apriori algorithm which is an extension of the Onto-Apriori algorithm. The extended algorithm is designed to improve the conventional algorithm's problem of comparing only identifiers of transaction items by reasoning transaction properties of the items which belong in the same category. We show how the mining algorithm works with a smartphone application recommender system based on our extended algorithm to clearly describe the procedures providing personalized recommendations. Further, our simulation results validate our analysis on the algorithm overhead, precision, and recall.

• Proceedings of the Safety Management and Science Conference
• /
• 2000.11a
• /
• pp.299-306
• /
• 2000

General definition of data mining is the knowledge discovery or is to extract hidden necessary information from large databases. Its technique can be applied into decision making, prediction, and information analysis through analyzing of relationship and pattern among data. One of the most important work is to find association rules in data mining. The objective of this paper is to find customer's trend using association rule from analysis of database and the result can be used as fundamental data for CRM(Customer Relationship Management). This paper uses Apriori algorithm and FoodMart data in order to find association rules.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.3
• /
• pp.15-21
• /
• 2008

This paper presents a new association mining algorithm that uses two phase sampling for shortening the execution time at the cost of precision of the mining result. Previous FAST(Finding Association by Sampling Technique) algorithm has the weakness in that it only considered the frequent 1-itemsets in trimming/growing, thus, it did not have ways of considering mulit-itemsets including 2-itemsets. The new algorithm reflects the multi-itemsets in sampling transactions. It improves the mining results by adjusting the counts of both missing itemsets and false itemsets. Experimentally, on a representative synthetic database, the algorithm produces a sampled subset of results with an increased accuracy in terms of the 2-itemsets while it maintains the same 1uality of the data set.

• Proceedings of the Korea Contents Association Conference
• /
• 2003.05a
• /
• pp.155-159
• /
• 2003

In order to compansate the problem of fragmentating data and disdaining small group of data in decision trees, a descriptive rule set discovery method is suggested. The principle of association rule finding algorithm is presented and a modified association nile finding algorithm for efficiency is applied to target database which has condition and decision attributes to see the effect of modification.

• The KIPS Transactions:PartC
• /
• v.10C no.6
• /
• pp.717-726
• /
• 2003

In this paper, we analyze the associated rule based deductive algorithm which creates the rules automatically for intrusion detection from the vast packet data. Based on the result, we also suggest the deductive algorithm which creates the rules of intrusion pattern fast in order to apply the intrusion detection systems. The deductive algorithm proposed is designed suitable to the concept of clustering which classifies and deletes the large data. This algorithm has direct relation with the method of pattern generation and analyzing module of the intrusion detection system. This can also extend the appication range and increase the detection speed of exiting intrusion detection system as the rule database is constructed for the pattern management of the intrusion detection system. The proposed pattern generation technique of the deductive algorithm is used to the algorithm is used to the algorithm which can be changed by the supporting rate of the data created from the intrusion detection system. Fanally, we analyze the possibility of the speed improvement of the rule generation with the algorithm simulation.

• Journal of Korea Multimedia Society
• /
• v.20 no.11
• /
• pp.1759-1767
• /
• 2017

Although Open API has been invigorated by advancements in the software industry, diverse types of malicious code have also increased. Thus, many studies have been carried out to discriminate the behaviors of malicious code based on API data, and to determine whether malicious code is included in a specific executable file. Existing methods detect malicious code by analyzing signature data, which requires a long time to detect mutated malicious code and has a high false detection rate. Accordingly, in this paper, we propose a method that analyzes and detects malicious code using association rule mining and an Naive Bayes classification. The proposed method reduces the false detection rate by mining the rules of malicious and normal code APIs in the PE file and grouping patterns using the DHP(Direct Hashing and Pruning) algorithm, and classifies malicious and normal files using the Naive Bayes.

• Journal of Korea Society of Industrial Information Systems
• /
• v.17 no.5
• /
• pp.9-17
• /
• 2012

Merchandise recommendations service based on electronic commerce has been actively studied and on service these days. By virtue of progress in IT industry, POS has been widely used even in small shops, but the merchandise recommendations service using POS has not been much facilitated compared with that of using electronic commerce. This paper proposes a merchandise recommendations service system using association analysis by applying data mining algorithm to POS sales data. This paper, also, suggests novel services such as annihilation rule and new rule, and ascending and descending rules. The analysis results are applied to the customers enabling to offer merchandise recommendations service. In addition, prompt responses against the changes in demands from customers are possible by identifying the annihilation rule and new rule, and ascending and descending rules, and providing the management with the rules as managerial decision making information.

• Journal of the Korea Society for Simulation
• /
• v.11 no.4
• /
• pp.15-23
• /
• 2002

As recently the network environment and application services have been more complex and diverse, there has. In this paper we introduce a scheme the extract useful information for network management by analyzing traffic data in user login file. For this purpose we use classification and association rule based on episode concept in data mining. Since login data has inherently time series characterization, convertible data mining algorithms cannot directly applied. We generate virtual transaction, classify transactions above threshold value in time window, and simulate the classification algorithm.