【摘要】：互聯(lián)網(wǎng)時(shí)代,數(shù)據(jù)量呈爆發(fā)式增長(zhǎng),對(duì)海量數(shù)據(jù)中有價(jià)值的信息的提取成為一種迫切需求。為了滿足這一需求,數(shù)據(jù)挖掘技術(shù)應(yīng)運(yùn)而生,并引起了越來越多的學(xué)者的關(guān)注。在這一背景下,諸多數(shù)據(jù)挖掘的理論觀點(diǎn)被不斷提出,基于微觀經(jīng)濟(jì)的數(shù)據(jù)挖掘框架就是其中之一。顧客導(dǎo)向的目錄分割問題是基于微觀經(jīng)濟(jì)觀點(diǎn)的數(shù)據(jù)挖掘在商業(yè)領(lǐng)域中的一個(gè)重要的研究課題,希望通過為企業(yè)設(shè)計(jì)適合的目錄發(fā)送給相應(yīng)的顧客,最大化被覆蓋的顧客的數(shù)量。本文在對(duì)顧客導(dǎo)向的目錄分割問題深入研究的基礎(chǔ)上,提出了新的目錄分割問題以及有效的目錄分割算法。在電子商務(wù)環(huán)境下,顧客的購(gòu)買行為受諸多因素影響,論文基于網(wǎng)上消費(fèi)者的購(gòu)買心理和行為,對(duì)加入利潤(rùn)約束的目錄分割問題進(jìn)行擴(kuò)展,提出考慮運(yùn)費(fèi)的雙目標(biāo)目錄分割問題,以最大化企業(yè)期望利潤(rùn)和被覆蓋顧客數(shù)為目標(biāo)。此外,基于顧客價(jià)值理論,考慮企業(yè)長(zhǎng)遠(yuǎn)效益,提出新的感知導(dǎo)向的目錄分割問題,以提高目錄商品的點(diǎn)擊率和轉(zhuǎn)化率。論文提出了一種改進(jìn)的協(xié)同遺傳算法,算法將種群分為兩個(gè)子種群,實(shí)現(xiàn)種群內(nèi)部進(jìn)化和種群之間的共同進(jìn)化,并通過最優(yōu)解集以及自適應(yīng)的變異算子的設(shè)計(jì),使算法的搜索過程跳出局部最優(yōu),加速算法的收斂。此外,利用合作交叉算子實(shí)現(xiàn)種群之間的信息交互,使得兩個(gè)子種群互相適應(yīng)、共同進(jìn)化�；谡鎸�(shí)的電商數(shù)據(jù),論文將提出的算法和memetic算法以及經(jīng)典的RBPF算法做對(duì)比分析,結(jié)果表明提出的算法能覆蓋更多的顧客,優(yōu)于前兩個(gè)算法�；谔岢鏊惴�,論文通過原型系統(tǒng)實(shí)現(xiàn)目錄定制的可視化。
[Abstract]:In the Internet era, the amount of data increases explosively, so it is an urgent need to extract valuable information from massive data. In order to meet this demand, data mining technology emerges as the times require, and has attracted more and more scholars' attention. Under this background, many theories of data mining have been put forward, and the framework of data mining based on microeconomics is one of them. The problem of customer-oriented catalog segmentation is an important research topic of data mining based on microeconomic viewpoint in the business field. It is hoped that it can be sent to the corresponding customers by designing suitable catalogs for enterprises. Maximize the number of customers covered. In this paper, a new directory segmentation problem and an effective directory segmentation algorithm are proposed based on the in-depth study of the customer oriented directory segmentation problem. In the electronic commerce environment, the customer's purchase behavior is affected by many factors. Based on the online consumer's purchase psychology and behavior, this paper extends the problem of catalog segmentation with profit constraints. This paper proposes a dual-objective catalog segmentation problem considering freight, which aims at maximizing the expected profit and the number of covered customers. In addition, based on the customer value theory and considering the long-term benefit of the enterprise, a new perception-oriented catalog segmentation problem is proposed to improve the click-through rate and conversion rate of catalog commodities. In this paper, an improved cooperative genetic algorithm is proposed. The algorithm divides the population into two sub-populations to realize the evolution within the population and the coevolution between the populations, and designs the optimal solution set and the adaptive mutation operator. The search process of the algorithm can jump out of the local optimum and accelerate the convergence of the algorithm. In addition, the cooperative crossover operator is used to realize the information exchange between the two populations, which makes the two subpopulations adapt to each other and co-evolve. Based on the real e-commerce data, the proposed algorithm is compared with the memetic algorithm and the classical RBPF algorithm. The results show that the proposed algorithm can cover more customers and is better than the first two algorithms. Based on the proposed algorithm, this paper realizes the visualization of catalog customization through the prototype system.
【學(xué)位授予單位】：安徽工業(yè)大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2017
【分類號(hào)】：F274

【參考文獻(xiàn)】

相關(guān)期刊論文 前5條

1 張麗萍;陳瑋;;面向壽險(xiǎn)客戶的目錄分割問題研究[J];計(jì)算機(jī)工程;2010年10期

2 徐秀娟;王U，

本文編號(hào)：2250138