We use cookies to give you the best experience possible. By continuing we’ll assume you’re on board with our cookie policy

See Pricing

What's Your Topic?

Hire a Professional Writer Now

The input space is limited by 250 symbols

What's Your Deadline?

Choose 3 Hours or More.
Back
2/4 steps

How Many Pages?

Back
3/4 steps

Sign Up and See Pricing

"You must agree to out terms of services and privacy policy"
Back
Get Offer

How Will Science and Techonology Changes Our Lives in Future

Hire a Professional Writer Now

The input space is limited by 250 symbols

Deadline:2 days left
"You must agree to out terms of services and privacy policy"
Write my paper

With the popularity of the Internet and the development of e-commerce, the recommendation system has gradually become an important component of ecommerce IT technology, and has drawn the more and more attention from esearchers and business people. However, most of the existing e-commerce systems use only part of the information available to make recommendations. With the development of the research, the new e-commerce recommendation system should take advantage of as much information as possible, collect various types of data, efficiently integrate multiple recommend technologies in order to provide more effective recommendation services.

At the same time, the expansion ability and the real-time requirement of the recommendation system are becoming more and more difficult to guarantee in a large-scale e-commerce recommendation system. So, the integration of multiple recommendation algorithms using various data and the real-time requirement are pressing problems in the development of e-commerce personalized service. In view of rich sources of e-commerce data, the key to problems can be nothing but the widespread application of the data mining technology and the establishment of a recommendation system model that can operate highly efficiently with multiple recommendation algorithms using various data.

Don't use plagiarized sources. Get Your Custom Essay on
How Will Science and Techonology Changes Our Lives in Future
Just from $13,9/Page
Get custom paper

4.2. System Architecture & Design
E-commerce personalized recommendation system and data mining Data mining is an uncommon process to extract the previously unknown and potentially useful information and knowledge from massive, incomplete, distributed, fuzzy and random data. This technology is widely used in classification, prediction and pattern recognition and so on. The biggest advantage of data mining technology for e-commerce is the massive data produced by the ecommerce conducts, which make just basis for data mining. At the same time, the e-commerce user information has many good characteristics such as the rich record, the good data type, research results that are easy to transfer. Therefore, data mining is very much necessary and applicable in the e-commerce recommendation system

The data mining system based on multi-agent
Agent generally refers to calculation entity with characteristics, such as independence, duration, sociality and acting as agent under certain environment. It has its own knowledge bases and reasoning mechanism, which can make a voluntary response to environment. Multi-Agent system (MAS) is a system composed of a lot of Agents, and generally these Agents exchange information through network infrastructure. In order to succeed in communicating, a certain Agent needs to cooperate and consult with other Agent [11]. In MAS, the ability of an individual Agent is limited, but multi-Agent can finish a lot of complicated tasks through cooperation [4]. MAS can improve enterprises capability to mine customer’s information effectively, which can economize much time and energy. Its basic thought is that a user corresponds to one Agent. When users search for goods on the e-commerce website, the management Agent will carry on pretreatment for data and establish user Agent, then mine the data and give information feedback to users finally, offer the individualized service. The frame of the system is designed as fig. 4.1 shows.

4.3. System Structure

.

4.4. The Functions of Agent in System

(1) Management Agent
It receive user’s request from the user’s graphic interface, then look over whether there is user’s information in Agent information storehouse. If it does not exist, establish a user Agent for this user, and provide a systematic serial number, and initialize the user’s model storehouse; if it does exist, activate the user Agent. And send users’ demand to user Agent, then give the information feedback excavated by mining Agent to the user. With collecting, managing and counting various kinds of information resources, it pretreats data collected in the e-commerce website, draws the relevant data, carries on two-justice analysis, dispels inconsistency, then transfers the treated data to mining Agent.

(2) User Agent
User Agent is responsible for communicating with management Agent and receiving and passing user’s order, to the mining Agent, finally analyses and coordinates the mining result transmitted by mining Agent to upgrade user’s model storehouse and preset information storehouse and submit to the management Agent at the same time. According to current user model storehouse and the order of customer, the user Agent provides feedbacks and studies, which make the information storehouse keep information alive all the time, eliminate that outdated, and offer the latest. The main arithmetic contained in the user Agent:

A. The work process of user Agent
On the stage of studying and carrying out on the backstage supporter, according to user’s past behavior, user Agent manages the knowledge in the user model storehouse, colligates task, analyzes user’s present operation intention, predicts its tendency, optimizes user model storehouse constantly, excavates out the most possible and frequent used information, and at the same time eliminates the outdated behavior record. While on the hanging up state, it releases the ample resources, but monitors the management Agent all the time. Once users send out order, it will enter the executive mode immediately.

B. Communication among multi-Agents
When users put forward a request, the corresponding user Agent looks over the preset information storehouse and finds whether there is corresponding service information, then carry on enlightenment mining to the data warehouse according to the preset user model, put forward cooperative request to other user Agent in case of the necessity to improve the mining efficiency and accuracy, which guarantees the effective sharing of the preset information storehouse to avoid resources waste. User Agent communicates with each other with KQML language. There is an action expression example based on KQML: User Agent i transfers request to user Agent j, and user Agent j receives the request then looks over his model storehouse, finally returns the result to user Agent i. The process is expressed as follows: (ask for

: sender Agent i
: receiver Agent j
: reply-with user j_model repository
: content (Userrequest(commodity material))
: language vb
: ontology commodity material)

(3) Mining Agent
The mining Agent is the analytic center of the whole Agent system. It mines the information registered in website, the daily visit record and shopping message of customer in website. Also, it excavates the data from user’s model storehouse according to the rule in the rule store- house. It mainly adopts two kinds of analytical methods, such as the related rule mining and cluster analysis [3].

(4) Information storehouse
There are two kinds of information storehouses in this system: One is user model storehouse, the other is Agent information storehouse. User model storehouse keeps user’s basic information, historical web-hit record and prediction of the customer’s possible future behaviors; furthermore it keeps ID number of other customers with similar interests and behaviors. While the Agent information storehouse keeps customer’s all materials in website, and classifies various users with the similar characteristics and behaviors according to cluster method, which benefits analysis and predicts a lot [11].

4.5. The Work Process of System

(1) Management Agent receives order submitted from graphic user interface. (2) Management Agent looks over the related information in the Agent information storehouse, activates or creates user Agent and transfers task to user Agent. (3) User Agent transfers information to mining Agent, then mining Agent carries on data mining for user’s task with various kinds mining methods, and provides feedbacks for user Agent. (4) User Agent communicates with other user Agent according to provided systematic serial number and transfers information to mining Agent. (5) User Agent assembles mining information to shape rule, updates user model and returns to management Agent. (6) Management Agent receives information, updates Agent information storehouse and feedbacks the mining result to user.

4.6. The Algorithm of Data Mining System based on Multi-agent

In B2C, online shopping is the most common one. In order to give a clear illustration, this paper takes an example of a customer buying the books of Cisco certification exam. This system mainly solves the problem of how to excavate out the purchase law utilizing Agent technology with the purpose to offer individual-customer oriented service. For online electronic shop, understanding the customer’s personal character and experience is a key factor to attract the long-term customer. The data in web daily record file reveals the customer’s browse character and improves customer’s interest in website.

4.6.1. The Application of Frequent-path (Frequent Path Algorithm)

Agent can get users’ search path from server cookie, which network topology diagram as fig.4.2 shows, analyze users’ interest according to the characteristics of website and the time user stays.

The user’s visit path in this website: {A->B->C ->D- >C ->B->A->E->F}. Here sets TP as the largest path (namely it is the biggest link page being called the heaviest forward path which begins from the beginning node until the last node having no backward) [4]. If P = ∑i-n i=1 ki .TP, i =1, 2…n, for i k is the hit number of i TP. Then the frequency of any heaviest forward path is f (TPi ) = ki. TPi / P [2]. If the value of f( TPi) is not less than the value stipulated by customer, we call this TPi as frequent visit path. We can understand which pages are interested by customer through fixing the frequent path, consequently it can better perfect the design to satisfy customer’s request. Therefore we can know that there are two largest paths in fig.4.2: (1) A->B->C->D (2) A ->E->F. We have adopted java language to realize the frequent path and the revised code is as follows: Void Frequentpath{

Cite this How Will Science and Techonology Changes Our Lives in Future

How Will Science and Techonology Changes Our Lives in Future. (2016, Jul 17). Retrieved from https://graduateway.com/how-will-science-and-techonology-changes-our-lives-in-future/

Show less
  • Use multiple resourses when assembling your essay
  • Get help form professional writers when not sure you can do it yourself
  • Use Plagiarism Checker to double check your essay
  • Do not copy and paste free to download essays
Get plagiarism free essay

Search for essay samples now

Haven't found the Essay You Want?

Get my paper now

For Only $13.90/page