• Proceedings of the Korean Institute of Intelligent Systems Conference
• /
• 2003.05a
• /
• pp.243-246
• /
• 2003

Data mining is a process of discovering useful patterns for decision making from an amount of data. It has recently received much attention in a wide range of business and engineering fields Classifying a group into subgroups is one of the most important subjects in data mining Tree-based methods, known as decision trees, provide an efficient way to finding classification models. The primary concern in tree learning is to minimize a node impurity, which is evaluated using a target variable in the data set. However, there are situations where multiple target variables should be taken into account, for example, such as manufacturing process monitoring, marketing science, and clinical and health analysis. The purpose of this article is to present several methods for measuring the node impurity, which are applicable to data sets with multiple target variables. For illustrations, numerical examples are given with discussion.

• Journal of the Korean Institute of Electrical and Electronic Material Engineers
• /
• v.21 no.11
• /
• pp.987-994
• /
• 2008

The optimal structure of the Tonpilz transducer was designed. First, the FE model of the transducer was constructed, that included all the details of the transducer which used practical environment. The validity of the FE model was verified through the impedance analysis of the transducer. Second, the resonance frequency, the sound pressure, the bandwidth, and the impulsive shock pressure of the transducer in relation to its structural variables were analyzed. Third, the design method of $2^n$ experiments was employed to reduce the number of analysis cases, and through statistical multiple regression analysis of the results, the functional forms of the transducer performances that could consider the cross-coupled effects of the structural variables were derived. Based on the all results, the optimal geometry of the Tonpilz transducer that had the highest sound pressure level at the desired working environment was determined through the optimization with the SQP-PD method of a target function composed of the transducer performance. Furthermore, for the convenience of a user, the automatic process program making the optimal structure of the acoustic transducer automatically at a given target and a desired working environment was made. The developed method can reflect all the cross-coupled effects of multiple structural variables, and can be extended to the design of general acoustic transducers.

• Journal of Korea Trade
• /
• v.24 no.6
• /
• pp.1-18
• /
• 2020

Purpose - This study investigates whether a listing effect exists in cross-border M&As and whether the effect can be attributed to the uncertainty of the GDP growth rate in the target firm's home country. We apply a joint variable analysis using M&A announcement data from the Korea Exchange (KRX), Shanghai Stock Exchange (SSE), and the Taiwan Stock Exchange (TWSE) from 2004 to 2013. We also conduct an event study using the measure of the uncertainty of the GDP growth rate (based on IMF statistics) in 55 target countries. Design/methodology - We measure the abnormal return (AR) using the market-adjusted model. We test the significance of the AR and the cumulative abnormal return (CAR) using a one-sample t-test. We examine the characteristics of the CARs depending on whether the target company is listed by applying a difference analysis using CAR as a test variable. In addition, we set CAR (-5, +5) as a dependent variable to identify the cause of the listing effect, and test both the financial characteristic variables of the acquirer and the collective characteristic variables of the merger as independent variables in the multiple regression analysis. Findings - First, we find the listing effect of cross-border M&As in the KRX, SSE, and TWSE, which represent the capital markets in Korea, China, and Taiwan, respectively. This listing effect persists during the global financial crisis and has a negative effect on the wealth of acquiring shareholders, especially when the target countries are emerging markets. Second, greater uncertainty regarding the target countries' economic growth in cross-border M&As has a negative effect on the wealth of acquiring firms' shareholders. Third, our empirical analysis demonstrates that the listing effect is attributable to the fact that firms listed in a target country with greater uncertainty of economic growth are more directly and greatly exposed to uncertain capital markets through stock markets, than are unlisted firms. Originality/value - This study is significant in that it presents a new strategic perspective in the study of cross-border M&As by demonstrating empirically that the listing effect is attributable to the uncertainty regarding the economic development of the target firms' home countries.

• The KIPS Transactions:PartC
• /
• v.18C no.2
• /
• pp.89-96
• /
• 2011

ASR is an excellent program security technique that protects various data memory areas without run-time overhead. ASR hides the addresses of variables from attackers by reordering variables within a data memory area; however, it can be broken by brute force attacks because of a limited data memory space. In this paper, we propose Multiple ASR to overcome the limitation of previous ASR approaches. Multiple ASR separates a data memory area into original and duplicated areas, and compares variables in each memory area to detect an attack. In original and duplicated data memory areas variables are arranged in the opposite order. This makes it impossible to overwrite the same variables in the different data areas in a single attack. Although programs with Multiple ASR show a relatively high run-time overhead due to duplicated execution, programs with many I/O operations such as web servers, a favorite attack target, show 40~50% overhead. In this paper we develop and test a tool that transforms a program into one with Multiple ASR applied.

• Journal of the Korean Data and Information Science Society
• /
• v.22 no.5
• /
• pp.877-884
• /
• 2011

The decision tree is a representative algorithm of data mining and used in many domains such as retail target marketing, fraud detection, data reduction, variable screening, category merging, etc. This method is most useful in classification problems, and to make predictions for a target group after dividing it into several small groups. When we create a model of decision tree with a large number of input variables, we suffer difficulties in exploration and analysis of the model because of complex trees. And we can often find some association exist between input variables by external variables despite of no intrinsic association. In this paper, we study on the removal method of unnecessary input variables using multiple external association rules. And then we apply the removal method to actual data for its efficiencies.

• Food Science and Biotechnology
• /
• v.16 no.4
• /
• pp.557-564
• /
• 2007

We prepared caviar analogs encapsulated by calcium-alginate gel membranes as a means to replace higher priced natural caviars. Processing the caviar analogs (beluga type) was optimized by response surface methodology with central composite design. Concentrations of sodium alginate ($X_1$) and $CaCl_2\;(X_2)$ were chosen as the independent variables. In order to compare characteristics of the caviar analogs with the natural caviar, sphericity ($Y_1$), diameter ($Y_2$), membrane thickness ($Y_3$), rupture strength ($Y_4$), rupturing deformation ($Y_5$), and sensory score ($Y_6$) were used as the dependent variables. The sphericity of the caviar analogs showed a similar value to that of natural caviar (over 94%) in the range of independent variables. Generally, the $CaCl_2$ concentration ($X_2$) affected all dependent variables to a greater extent than the sodium alginate concentration ($X_l$), For the multiple response optimization of the 5 dependent variables ($Y_1,\;Y_2,\;Y_4,\;Y_5$, and $Y_6$), the desirability function was defined as the following conditions: target values ($Y_1\;=\;100%,\;Y_2\;=\;3.0\;mm,\;Y_4\;=\;1,470\;g,\;Y_5\;=\;1.1\;mm,\;and\;Y_6\;=\;10\;points$). Membrane thickness ($Y_3$) was eliminated from the dependent variables for multiple response optimization because it could not be measured with an image analyzer. The values of the independent variables as evaluated by multiple response optimization were $X_1\;=\;-0.093$ (78%) and $X_2\;=\;-0.322$ (1.07%), respectively.

• Journal of Korean Society for Quality Management
• /
• v.23 no.3
• /
• pp.84-92
• /
• 1995

A method is proposed for the simultaneous optimization of several response functions that depend on the same set of controllable variables and are adequately represented by a response surface model (polynomial regression model) with the same degree and with constraint that the individual responses have the target values. First, the multiple responses data are checked for linear dependencies among the responses by eigenvalue analysis. Thus a set of responses with no linear functional relationships is used in developing a function that measures the distance estimated responses from the target values. We choose the optimal condition that minimizes this measure. Also, under the different degree of importance two step procedures are proposed.

• Nuclear Engineering and Technology
• /
• v.56 no.6
• /
• pp.2266-2273
• /
• 2024

A machine learning platform is proposed for the diagnosis of a severe accident progression in a nuclear power plant. To predict the key parameters for accident management including lost signals, a long short term memory (LSTM) network is proposed, where multiple accident scenarios are used for training. Training and test data were produced by MELCOR simulation of the Fukushima Daiichi Nuclear Power Plant (FDNPP) accident at unit 3. Feature variables were selected among plant parameters, where the importance ranking was determined by a recursive feature elimination technique using RandomForestRegressor. To answer the question of whether a reduced order ML model could predict the complex transient response, we performed a systematic sensitivity study for the choices of target variables, the combination of training and test data, the number of feature variables, and the number of neurons to evaluate the performance of the proposed ML platform. The number of sensitivity cases was chosen to guarantee a 95 % tolerance limit with a 95 % confidence level based on Wilks' formula to quantify the uncertainty of predictions. The results of investigations indicate that the proposed ML platform consistently predicts the target variable. The median and mean predictions were close to the true value.

• Journal of Korean Society for Quality Management
• /
• v.27 no.2
• /
• pp.218-236
• /
• 1999

The purpose of parameter design is to determine optimal settings of design parameters of a product or a process such that the performance characteristics of a product exhibit small variabilities around their target values. Taguchi made significant contributions in this area. However, his analysis of the problem focused on only one performance characteristic or response, although in product and process design, multiple characteristics are more common. The critical problem in dealing with multiple characteristics is how to compromise the conflict among the selected levels of the design parameters for each individual characteristic. In this paper, Methodology using SN ratio optimized by univariate technique is proposed and a parameter design procedure to achieve the optimal balance among several different response variables is developed. Existing case studies are solved by the proposed method and the results are compared with ones by the sum of SN ratios, the expected weighted loss, the desirability function, and EXTOPSIS model.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.4
• /
• pp.386-390
• /
• 2003

Data mining is a process of discovering useful patterns for decision making from an amount of data. It has recently received much attention in a wide range of business and engineering fields. Classifying a group into subgroups is one of the most important subjects in data mining. Tree-based methods, known as decision trees, provide an efficient way to finding the classification model. The primary concern in tree learning is to minimize a node impurity, which is evaluated using a target variable in the data set. However, there are situations where multiple target variable should be taken into account, for example, such as manufacturing process monitoring, marketing science, and clinical and health analysis. The purpose of this article is to present some methods for measuring the node impurity, which are applicable to data sets with multivariate target variables. For illustration, a numerical cxample is given with discussion.