USST_Arts_112400425 Web Mashup中数据源的组织及管理技术

VIP免费

3.0 牛悦 2024-11-11 5 4 4.13MB 47 页 15积分

侵权投诉

Web Mashup 中数据源的组织及管理技

术

摘要

随着科技的发展，人们对于 Web 服务有着个性化、多样化的要求，传统单一

的Web 服务难于满足用户个性化的需求，通常需要通过几个 Web 服务才能解决。

开发者利用 Mashup 技术，根据用户的需求聚合多个 Open API 服务形成一个新的

应用。随着 Open API 数量的快速增长，开发者从中找到所需的 Open API 变得很困

难。如何有效地组织和管理这些 Open API 成为一个迫切需要解决的问题。

为了有效地管理 Open API，首先，本文提出了一个 Open API 描述模型，该

模型是基于 Open API 对外接口信息的参数化描述。然后在非度量空间下利用聚类

算法对 Open API 进行分类；由于 Open API 之间的相似性具有非度量空间的特性，

因此计算结果更加精确。同时，本文提出一个 BKCR 推荐算法，它能根据开发者

选择的 Open API，在系统范围内搜索能形成 Mashup 应用的和相近的 Open API。

作为 Open API 描述模型和 BKCR 算法的实现，本文搭建了一个 Open API 数

据源的组织管理系统。该系统不仅可很方便地对 Open API 进行新增、更新、删除、

分类等管理维护，而且还可以使用推荐模块，为开发者推荐优质的 API。

本文贡献如下：

1. 提出一个 Open API 描述模型，该模型是一个十元向量，它除了描述 Open

API 数据质量、使用协议、数据格式外，还能反映其在系统中的状态和结合 Open

API 非度量空间的特性。因而，它能够更好地代表 Open API，为后续的聚类、推荐

等做准备。

2. 研究 Open API 间相似性的计算方法，现实中 Open API 的每个维度重要性，

不是完全不等的，而且不同对象之间相似维度也不一致。结合非度量空间的性质。

本文提出一个加权的非度量空间相似性计算方法，它可以更准确地描述两 Open

API 间的相似性。

3. 提出一个 BKCR 推荐算法，它结合了传统基于项目的推荐算法和关联规

则。根据用户选择 Open API 的关键词信息，计算它和关联规则前件的相似性和推

荐算法的推荐结果，将相似度满足设置阈值的结果返回。因此， BKCR 具有更广

泛的适用性。

4. 最后通过一系列的实验，验证本文提出的描述模型和推荐算法的性能，判

断其能有效地管理数据源。首先通过对比 Metric space 和Non-metric space 中相似

度计算准确率，再者就是比较一般的推荐算法和 BKCR 算法性能的比较。结果表

明，本文所提模型和 BKCR 相比一般的模型和推荐算法具有更好的效果。

关键词：Open API 描述模型推荐算法聚类算法 Mashup

ABSTRACT

With the development of science and technology, the development of web services

is personalized and diversified. Normally, it takes more than one traditional web service

to meet the users’ personalized needs. Developers use Mashup technology to aggregate

multiple Open API services into a new application according to the needs of customers.

Since the number of Open API grows rapidly, it is more and more difficult for

developers to find the most appropriate Open API services. How to organize and

manage these Open API services effectively becomes an urgent problem for people to

solve.

In order to manage Open API effectively, first of all, this paper proposes an Open

API description model. This model is a parametric description for the external interface

information of Open API. Second, this paper takes the clustering algorithm of non-

metric space to classify Open API. Because the similarity between Open API have the

properties of non-metric space, using the clustering algorithm of non-metric space will

make the results more accurate. Finally, this paper proposes a BKCR recommendation

algorithm, which can recommend a list of highly similar and Mashup-applied Open API

according to the developers’ choice.

This paper has built a data source management system of Open API to implement

the description model of Open API and BKCR recommendation algorithm. This system

can not only maintaining Open API easily, but also can use recommended algorithm to

search the Mashup-applied API.

The contributions of this paper are as follows:

1. An Open API description model is proposed. This model is a ten-element

vector, and it can not only describe the Open API’ quality of data, the protocol used and

data format, but also can combine the properties of non-metric space and reflect Open

API’s state in this system. In a word, it can represent Open API better, and prepares for

the subsequent clustering and recommendation.

2. Study the calculation methods of similarities between Open API. The

importance of each dimension in Open API is completely different in real world, and the

similar dimensions between different objects are not fixed. This paper proposes a

weighted similarity calculation method, which is based on the property of non-metric

space, and it can describe the similarity between two API more accurately.

3. This paper proposes the BKCR recommendation algorithm, which combines

traditional recommendation algorithm and association rules. According to the key

information of Open API selected by users, the BKCR can calculate the similarities of

association rule antecedent and the result of the BKCR recommends. Obviously, it has

better extensive applicability.

4. Finally, the description model and BKCR recommendation algorithm in this

paper has been validated with good performance and managing the data source

effectively through a series of experiments. Firstly, contrasting the accuracy of

similarity-calculation in metric space and non-metric space; secondly, contrasting

effective and efficient between the normal recommendation and the BKCR. The results

show that the proposed model and BKCR have better performance compared to the

general model and recommendation algorithms.

Keywords: Open API, API model, recommendation algorithm,

clustering algorithm, Mashup

目  录
中文摘要
ABSTRACT
第一章 绪论..............................................................................................................1
1 研究背景.........................................................................................................1
2 国内外研究现状.............................................................................................3
2.1 API 描述模型...............................................................................................3
2.2 Mashup 应用中 API 间的关系....................................................................4
3 研究内容与难点.............................................................................................5
4 论文组织结构.................................................................................................6
第二章 Web Mashup 概述及关键技术.....................................................................7
1 Mashup 概述.......................................................................................................7
1.1 Mashup 研究现状........................................................................................7
1.2 Mashup 应用框架........................................................................................8
2 Mashup 应用分类...............................................................................................9
3 Mashup 相关技术和协议..................................................................................11
3.1 AJAX..........................................................................................................11
3.2 通信协议....................................................................................................12
3.3 数据类型....................................................................................................13
4 关键技术....................................................................................................14
4.1 聚类算法....................................................................................................14
4.2 匹配推荐算法............................................................................................16
5 本章小结...........................................................................................................17
第三章Web API 的描述模型.................................................................................18
1 Web API 的表示方法........................................................................................18
1.1 目前 API 表示法........................................................................................18
1.2 API 动态属性表示法.................................................................................19
2 Web API 间的非度量空间特性........................................................................20
2.1 非度量空间................................................................................................20
2.2 Web API 非度量空间的表现.....................................................................21
2.3 常用非度量空间距离计算方法................................................................22
3 Web API 安全性................................................................................................23
3.1 Web API 安全技术.....................................................................................23
3.2 安全性与推荐排序....................................................................................24
4 本章小结...........................................................................................................24
第四章Web API 推荐算法.....................................................................................25
1 聚类分析...........................................................................................................25
2 关联规则...........................................................................................................26
2.1 关联规则基础概念....................................................................................26
2.2 关联规则产生一般模型............................................................................27
2.3 关联规则与推荐算法................................................................................28
3 基于项目的协同推荐算法...............................................................................28

4 基于关键字的推荐算法...................................................................................29
4.1 相似性计算方法........................................................................................29
4.2 BKCR 推荐算法........................................................................................30
5 BKCR 算法的性能分析...................................................................................31
6 本章小结...........................................................................................................31
第五章Web API 组织管理系统.............................................................................32
1 API 管理系统设计............................................................................................32
2 Web API 管理模块............................................................................................33
3 Web API 推荐模块............................................................................................34
3.1 Web API 推荐方案....................................................................................34
3.2 Web API 推荐流程....................................................................................35
3.3 Web API 推荐实现步骤............................................................................37
4 实验结果及相关分析.......................................................................................39
4.1 实验设置....................................................................................................39
4.2 实验结果及分析........................................................................................40
5 本章小结...........................................................................................................45
第六章总结与展望................................................................................................46
参考文献.....................................................................................................................48

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

15 积分 4人已下载

立即下载 VIP免费下载

摘要：

WebMashup中数据源的组织及管理技术摘要随着科技的发展，人们对于Web服务有着个性化、多样化的要求，传统单一的Web服务难于满足用户个性化的需求，通常需要通过几个Web服务才能解决。开发者利用Mashup技术，根据用户的需求聚合多个OpenAPI服务形成一个新的应用。随着OpenAPI数量的快速增长，开发者从中找到所需的OpenAPI变得很困难。如何有效地组织和管理这些OpenAPI成为一个迫切需要解决的问题。为了有效地管理OpenAPI，首先，本文提出了一个OpenAPI描述模型，该模型是基于OpenAPI对外接口信息的参数化描述。然后在非度量空间下利用聚类算法对OpenAPI进行分类...

展开>> 收起<<

USST_Arts_112400425 Web Mashup中数据源的组织及管理技术.doc

共47页,预览5页

还剩页未读，继续阅读

USST_Arts_112400425 Web Mashup中数据源的组织及管理技术

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

推荐作者

热门标签

举报选择: