應(yīng)用于電子商務(wù)環(huán)境的商業(yè)模式挖掘和預(yù)測方法（英文）
發(fā)布時間：2019-04-12 20:24
【摘要】：目的:隨著電子商務(wù)的發(fā)展,商品交易形式發(fā)生了翻天覆地的變化。在這種新型虛擬交易平臺上,消費者快速而又廣泛地瀏覽、購買、評價各種價廉物美的商品,而商品也同樣地進(jìn)行著產(chǎn)品創(chuàng)新、營銷拓展、物流改善。所以,消費者和商品之間早已不是簡單的一對一直接買賣關(guān)系,而是消費者和消費者之間的社交網(wǎng)絡(luò),消費者和商品之間的交易網(wǎng)絡(luò)構(gòu)成電子商務(wù)中最重要的物質(zhì)載體。于是,對于商品銷量預(yù)測這一傳統(tǒng)經(jīng)濟(jì)問題,在電子商務(wù)的大環(huán)境下有了新型的研究意義。為了能有效和深入地研究在電子商務(wù)環(huán)境下的商品銷量發(fā)展模式,本文的目的在于提供一種應(yīng)用于電子商務(wù)的商品銷量預(yù)測算法,該算法注重于從消費者社會影響力分析入手,更好地適應(yīng)真實電子商務(wù)環(huán)境中的銷量預(yù)測需求。創(chuàng)新點:首先,本文算法不僅考慮到消費者自身的特征,同時還考慮到存在于消費者之間的社會影響力,考慮到在真實的電子商務(wù)中,消費者之間傳遞商品的價格信息或評價信息十分便捷,因而本文算法很好地切合了實際的應(yīng)用環(huán)境。其次,本文算法定義了交易環(huán)境中的兩種社會影響力,即"同一商品中消費者互相作用產(chǎn)生的影響力"和"不同商品之間消費者互相作用產(chǎn)生的影響力",分別考慮到單一商品的交易環(huán)境和多個商品互相作用的交易環(huán)境中消費者行為,其中以上兩種社會影響力都是由真實消費者社交網(wǎng)絡(luò)分析提煉得來的,使得本算法更加切合真實交易網(wǎng)絡(luò)的內(nèi)在結(jié)構(gòu)。方法:本文算法將商品銷量分為主體部分和噪聲部分,很好地模擬了真實交易環(huán)境中,商品銷量的構(gòu)成是受多成分影響的。并且在預(yù)測模型中,對主體部分和噪聲部分分別設(shè)置了不同的約束條件,具體為要求商品銷量的主體部分在時間上應(yīng)該保持平滑性,并要求商品銷量的噪聲部分是稀疏的,以上兩個約束很好地反映了真實交易環(huán)境中商品銷量的變化形式。結(jié)論:本文研究電子商務(wù)環(huán)境下商品銷量的發(fā)展模式,并提出描述消費者之間關(guān)系的兩種社會影響力網(wǎng)絡(luò)。將此社會影響力網(wǎng)絡(luò)整合入商品銷量構(gòu)成模型中,最后提出對這些商品銷量的預(yù)測算法。特別是通過在真實的數(shù)據(jù)環(huán)境中(阿里巴巴女裝數(shù)據(jù))進(jìn)行算法測試,并結(jié)合與傳統(tǒng)銷量預(yù)測算法的比較,展示在復(fù)雜數(shù)據(jù)環(huán)境下本算法的有效性。
[Abstract]:Objective: with the development of e-commerce, the form of commodity trade has changed dramatically. On this new virtual trading platform, consumers browse, purchase and evaluate all kinds of cheap and beautiful goods quickly and extensively, and the goods also carry on product innovation, marketing expansion and logistics improvement. Therefore, the relationship between consumers and goods is not a simple one-to-one direct trading relationship, but a social network between consumers and consumers, and the transaction network between consumers and goods constitutes the most important material carrier in e-commerce. Therefore, for the traditional economic problem of commodity sales forecast, it has a new research significance under the environment of e-commerce. In order to effectively and deeply study the development mode of commodity sales under the environment of e-commerce, the purpose of this paper is to provide an algorithm for forecasting the sales volume of goods applied to e-commerce. The algorithm focuses on the analysis of consumers' social influence. Better adapt to the real e-commerce environment sales forecast demand. Innovation: first of all, this algorithm not only takes into account the characteristics of consumers, but also considers the social influence between consumers, taking into account that in real e-commerce, It is very convenient for consumers to transfer the price information or evaluation information of goods, so the algorithm in this paper fits well with the practical application environment. Secondly, the algorithm defines two kinds of social influence in the trading environment, that is, "the influence of consumer interaction in the same commodity" and "the influence of consumer interaction between different commodities". Considering the trading environment of a single commodity and the behavior of consumers in a trading environment in which multiple goods interact, the above two kinds of social influence are extracted from the analysis of real consumer social networks. It makes this algorithm more suitable to the internal structure of the real trading network. Methods: this algorithm divides the product sales into the main part and the noise part, and well simulates the real trading environment, the composition of the commodity sales is affected by the multi-component. And in the prediction model, different constraints are set on the main part and the noise part, specifically, the main part of the commodity sales should be smooth in time, and the noise part of the product sales is sparse, and the main part of the product sales should be smooth in time, and the noise part of the product sales should be sparse. The above two constraints well reflect the changing form of commodity sales in the real trading environment. Conclusion: this paper studies the development model of commodity sales in e-commerce environment, and puts forward two kinds of social influence networks which describe the relationship between consumers. This network of social influence is integrated into the model of commodity sales, and finally, the prediction algorithm of these commodity sales is put forward. Especially, the algorithm is tested in the real data environment (Alibaba women's dress data) and compared with the traditional sales forecasting algorithm to show the effectiveness of the algorithm in the complex data environment.
【作者單位】： Department
【基金】：supported by the National Basic Research Program(973)of China(No.2012CB316400) Zhejiang University-Alibaba Financial Joint Lab,Zhejiang Provincial Engineering Center on Media Data Cloud Processing and Analysis,China the US National Science Foundation(No.CCF-1017828)
【分類號】：F724.6;TP301.6
[Abstract]:Objective: with the development of e-commerce, the form of commodity trade has changed dramatically. On this new virtual trading platform, consumers browse, purchase and evaluate all kinds of cheap and beautiful goods quickly and extensively, and the goods also carry on product innovation, marketing expansion and logistics improvement. Therefore, the relationship between consumers and goods is not a simple one-to-one direct trading relationship, but a social network between consumers and consumers, and the transaction network between consumers and goods constitutes the most important material carrier in e-commerce. Therefore, for the traditional economic problem of commodity sales forecast, it has a new research significance under the environment of e-commerce. In order to effectively and deeply study the development mode of commodity sales under the environment of e-commerce, the purpose of this paper is to provide an algorithm for forecasting the sales volume of goods applied to e-commerce. The algorithm focuses on the analysis of consumers' social influence. Better adapt to the real e-commerce environment sales forecast demand. Innovation: first of all, this algorithm not only takes into account the characteristics of consumers, but also considers the social influence between consumers, taking into account that in real e-commerce, It is very convenient for consumers to transfer the price information or evaluation information of goods, so the algorithm in this paper fits well with the practical application environment. Secondly, the algorithm defines two kinds of social influence in the trading environment, that is, "the influence of consumer interaction in the same commodity" and "the influence of consumer interaction between different commodities". Considering the trading environment of a single commodity and the behavior of consumers in a trading environment in which multiple goods interact, the above two kinds of social influence are extracted from the analysis of real consumer social networks. It makes this algorithm more suitable to the internal structure of the real trading network. Methods: this algorithm divides the product sales into the main part and the noise part, and well simulates the real trading environment, the composition of the commodity sales is affected by the multi-component. And in the prediction model, different constraints are set on the main part and the noise part, specifically, the main part of the commodity sales should be smooth in time, and the noise part of the product sales is sparse, and the main part of the product sales should be smooth in time, and the noise part of the product sales should be sparse. The above two constraints well reflect the changing form of commodity sales in the real trading environment. Conclusion: this paper studies the development model of commodity sales in e-commerce environment, and puts forward two kinds of social influence networks which describe the relationship between consumers. This network of social influence is integrated into the model of commodity sales, and finally, the prediction algorithm of these commodity sales is put forward. Especially, the algorithm is tested in the real data environment (Alibaba women's dress data) and compared with the traditional sales forecasting algorithm to show the effectiveness of the algorithm in the complex data environment.
【作者單位】： Department
【基金】：supported by the National Basic Research Program(973)of China(No.2012CB316400) Zhejiang University-Alibaba Financial Joint Lab,Zhejiang Provincial Engineering Center on Media Data Cloud Processing and Analysis,China the US National Science Foundation(No.CCF-1017828)
【分類號】：F724.6;TP301.6
【共引文獻(xiàn)】
相關(guān)期刊論文 前10條
1 曾伏娥;代婷婷;朱妮亞;;網(wǎng)絡(luò)社區(qū)成員回應(yīng)社區(qū)廣告的社會性影響因素研究[J];管理學(xué)報;2013年08期
2 王永貴;馬雙;;虛擬品牌社區(qū)顧客互動的驅(qū)動因素及對顧客滿意影響的實證研究[J];管理學(xué)報;2013年09期
3 范鈞;付沙沙;湯錦旦;;虛擬品牌社區(qū)互動對顧客公民行為的影響[J];商業(yè)研究;2014年02期
4 李雪欣;包文莉;夏天;;虛擬品牌社區(qū)顧客動機(jī)對產(chǎn)品創(chuàng)新行為的影響[J];東北大學(xué)學(xué)報(社會科學(xué)版);2014年05期
5 胡磊;高迎;;微博社區(qū)成員參與動機(jī)實證研究[J];北京大學(xué)學(xué)報(自然科學(xué)版);2014年05期
6 胡磊;王瓊;范p，
本文編號：2457365
本文鏈接：http://sikaile.net/guanlilunwen/yingxiaoguanlilunwen/2457365.html
最近更新
教材專著
