集合住宅修繕資料之轉碼模式建構

Develop a Transcoding Model for Maintenance Records of Residential Buildings

10.29468/PMR.201109.0321

張智元(Chih-Yuan Chang)；張宇爵(Yu-Chueh Chang)

物業管理 ； 文字探勘 ； 維護 ； 轉碼 ； 編碼 ； Property Management ； Text Mining ； Maintenance ； Transcoding ； Coding

物業管理學會論文集

第5屆（2011 / 09 / 03）

321 - 328

繁體中文

國內物業管理公司少數有自己一套編製修繕資料的系統，但各公司所編製的代號或代碼多數不一致，大部分皆未考量修繕事件之編碼問題，導致日後在調查或分析建物損壞事件的資訊編碼時，無法有效應用歷史修繕資料，甚至因無一致性編碼基礎而無法跨平台產生資訊交流的應用價值。因此本研究所設計之建築疾病轉碼系統（Building Disease Transcoding System, BDTS），就是為了因應前述問題所研發之工具，讓修繕紀錄之疾病編碼具有結構化的特性與分類，以助於大量修繕資料之統分析交換與應用。研究中透過建築疾病特性及修繕資料特性之關聯，並結合建築疾病分類（BDC）編碼、同義詞與文字探勘技術（Text Mining），建置建築疾病轉碼模型（Building Disease Transcoding Model, BDTM），進而使用程式語言（Visual Basic, VB）研發出建築疾病轉碼系統（BDTS），BDTS已能準確的從修繕資料內容中萃取出關鍵資訊，直接轉出設施設備之構件碼與建築物損壞種類之編碼，並且經由6000筆的修繕資料測試驗證，目前轉碼成功率95%，由此可得知建築疾病轉碼系統之可用性。

Currently, few of the domestic Property Management Companies have adopted its own edit system to precede maintenance records. But these systems that edit numbers or coding were varied in various companies, and often does not take coding problem of renovated events into consideration. As result, as one required to investigate or analyze the data coding of damaged events of building structures, it often were unable to extract historical data efficiently. Moreover, such inconsistent coding-foundation has caused inability to generate data exchange on the platform, which there was no applicable value for these coding. Therefore, this study has designed the Building Disease Transcoding System (BDTS) that was to correspond with above illustrated problems. Such system was able to integrate the disease coding of renovated record to become more structural on its characteristic and classification, which was able to assist on statistical analyses exchange and application on mass amount of renovated data.Through the relationship between building disease characteristic and maintenance record characteristic; with support of BDC coding and synonym word-base and integration of Text Mining and Building Disease Transcoding Model (BDTM), and further utilization of Visual Basic (VB) in order to develop BDTS. The BDTS was able to extract key data from maintenance record accurately, and convert the coding of facility component and building structure damaged classification directly. Such conversion has been evaluated from 6000 entries of maintenance data with 95% of successful rate, which has shown the applicableness of BDTS.