專利授權區國立清華大學國際產學營運總中心 Operations Center for Industry Collaboration

搜尋專利授權區

關鍵字

» 新增關鍵字

選單

專利授權區

專利授權區
專利名稱(中)	回歸模型之最佳化方法及其系統
專利名稱(英)	OPTIMIZE METHOD AND SYSTEM OF REGRESSION MODEL
專利家族	中華民國：I912815 美國：US-2026-0003930-A1(公開號)
專利權人	國立清華大學 100.00%
發明人	葉維彰
技術領域	資訊工程,工業工程

專利摘要(中)
一種回歸模型之最佳化方法，包含依據簡化群體生成規則產生參數組，參數組包含複數區間值及複數模型代號。此些模型代號對應複數個回歸模型，且此些回歸模型的類別相異。依據遞增次序排列此些區間值。根據此些區間值依序將資料集的複數資料分為複數群體。此些群體分別對應此些模型代號。根據此些群體對應的此些模型代號對此些群體中的此些資料計算而產生預測結果及預測結果的適應值。依據參數組對應的適應值更新資料庫的最佳適應值，重複執行上述步驟，直到複數參數組的數量等於一預設數量。藉此，降低回歸模型計算結果的殘差。

專利摘要(中)

一種回歸模型之最佳化方法，包含依據簡化群體生成規則產生參數組，參數組包含複數區間值及複數模型代號。此些模型代號對應複數個回歸模型，且此些回歸模型的類別相異。依據遞增次序排列此些區間值。根據此些區間值依序將資料集的複數資料分為複數群體。此些群體分別對應此些模型代號。根據此些群體對應的此些模型代號對此些群體中的此些資料計算而產生預測結果及預測結果的適應值。依據參數組對應的適應值更新資料庫的最佳適應值，重複執行上述步驟，直到複數參數組的數量等於一預設數量。藉此，降低回歸模型計算結果的殘差。

專利摘要(英)
An optimized method of regression model includes generating a parameter set according to a Simplification Swarm Optimization (SSO) rule. The parameter set includes a plurality of threshold values and a plurality of model codes. The model codes are corresponding to a plurality of regression models, and the type of the regression models are different. The threshold values are arranged according to an increment sequence. A plurality of data is divided into a plurality of groups in sequence according to the threshold values, and the groups are corresponding to the model codes, respectively. The data of the groups are calculated to generate a predicting result and a fitness value of the predicting result according to the model codes corresponding to the groups. A best fitness value in a database is updated according to the fitness value corresponding to the parameter set, and the aforementioned steps are executed repeatedly until a number of a plurality of the parameter sets equal to a predetermined value. Thus, the optimize method of regression model of the present disclosure can reduce the residual error of the calculate result of the regression model.

專利摘要(英)

An optimized method of regression model includes generating a parameter set according to a Simplification Swarm Optimization (SSO) rule. The parameter set includes a plurality of threshold values and a plurality of model codes. The model codes are corresponding to a plurality of regression models, and the type of the regression models are different. The threshold values are arranged according to an increment sequence. A plurality of data is divided into a plurality of groups in sequence according to the threshold values, and the groups are corresponding to the model codes, respectively. The data of the groups are calculated to generate a predicting result and a fitness value of the predicting result according to the model codes corresponding to the groups. A best fitness value in a database is updated according to the fitness value corresponding to the parameter set, and the aforementioned steps are executed repeatedly until a number of a plurality of the parameter sets equal to a predetermined value. Thus, the optimize method of regression model of the present disclosure can reduce the residual error of the calculate result of the regression model.

聯絡資訊
承辦人姓名	黃允恬
承辦人電話	(03)571-5131#62305
承辦人Email	yuntian@mx.nthu.edu.tw