


摘 ?要: 本文以量子遺傳SVM為核心,構建數學模型,對公共圖書館每日讀者流通人次(借還書人次)進行預測。模型以上海市嘉定區圖書館為數據實例,考察各種影響因素后,以若干量化特征數據為輸入,每日讀者流通人次為輸出目標,在此基礎之上成功地訓練并獲取了預測模型。實驗表明,在全程貼近實際目標預測系統構建的指導思想下(適用、穩定、準確),此方法建立的相應短期預測系統取得了較好的預測效果,系統的應用將有效地提升圖書館智慧服務的層級。
關鍵詞: 量子遺傳算法;SVM;公共圖書館;智慧服務;每日讀者流通人次;量化特征
中圖分類號: TP181;G251.5 ? ?文獻標識碼: A ? ?DOI:10.3969/j.issn.1003-6970.2019.12.042
本文著錄格式:鄭戍嘉. 基于量子遺傳SVM的公共圖書館每日讀者流通人次預測模型[J]. 軟件,2019,40(12):188194
Prediction System of “Public Library Daily Readers Throughput”
Based on Quantum Genetic SVM
ZHENG Shu-jia
(Jiading District Library of Shanghai, Shanghai 201800, China)
【Abstract】: Based on the Quantum Genetic SVM, this paper constructs mathematical models to predict “public library daily readers throughput”, i.e. the amount of daily library users who borrow/return books. The model is taking Jiading District Library of Shanghai as data source. After investigating and analyzing various potential influencing factors, the model takes several quantitative characteristic data as input and sets “public library daily throughput” as target output, on the basis of which the prediction model is trained and successfully obtained. Experiments show that the corresponding short-term forecasting system established by this method has achieved good forecasting results under the guidance of the construction of the forecasting system which is close to the actual target (applicable, stable and accurate). The application of the system will effectively improve the level of library intelligent service.
【Key words】: Quantum genetic algorithm; SVM; Public library; Intelligent service; Daily readers throughput; Quantitative characteristics
0 ?引言
公共圖書館的每日讀者流通人次(借還書人次),是一個極有業務參考價值的指標式數據,它是在多種復雜因素共同作用下產生的。
目前國內對圖書館流通人次(冊次)時間序列的預測,粒度多集中在月度、學期的時間跨度上,而預測每日流通人次的嘗試,則較為少見,有個別文獻使用基于某型神經網絡的算法模型對高校圖書
館的每日流通人次進行了成功的短期預測,但對公共圖書館的每日流通人次的高精度預測模型則未曾見諸公開發表。
高校圖書館與公共圖書館流通人次的時間序列模型,雖看則相似,但實際存在著很大的不同。這主要是因為,兩者在流通模式、讀者特征、借閱行為模式、服務內容、交通方式乃至天氣因素影響等方面存在著較大差異,簡單地套用模型將會導致不甚理想的結果。……