| Title: | 软件成本最优线性组合估算方法与应用研究 |
| Author: | 李奇
|
| Issued Date: | 2008-05-29
|
| Degree Grantor: | 中国科学院软件研究所
|
| Place of Degree Grantor: | 软件研究所
|
| Degree Level: | 博士
|
| Keyword: | 最优线性组合
; 软件成本估算
; 组合预测
|
| Alternative Title: | Estimate Software Cost Using Optimal Linear Combination Method with a Case Study
|
| Abstract: | 软件成本估算的方法和技术众多且各有千秋，没有一种方法被证明在各种情况下的估算都是最精确的。组合估算是一种非常有效的估算方法，它避免了去选择一种单一的最优的估算方法，而是充分利用各种估算方法提供的信息来达到提高估算精度的目的。
在这篇文章中，我们首先将系统的组合预测方法介绍到软件成本估算领域，然后我们分类介绍了当今流行的软件成本估算方法及工具，分析它们各自的特点，这是我们进行组合估算的前提。同时我们介绍了怎样选择互补子模型进行组合的方法。
接着我们介绍了最具代表性的组合估算方法：最优线性组合（Optimal Linear Combining）估算方法并且用这种方法在一个真实数据集上进行了实例研究。结果表明组合不同的软件成本估算方法能够显著提高软件成本估算的精度和稳定性，即使是估算效果“较差”的技术或者方法也能为组合估算精度的提高做出贡献。
最后我们进一步讨论了影响OLC估算精度的因素，进行了详细的数据敏感性分析，提出了一种识别影响OLC精度提高的异常点的方法并进行了有效性验证。
通过本文的研究，证明了组合估算方法在实际应用中确实能够充分利用各种方法的估算信息提高软件成本的估算精度，组合预测已经成为预测领域未来发展方向之一，然而软件成本估算在这方面研究却几乎还是空白，我们希望更多的研究者投入到这方面的研究上来，共同进行这项充满挑战但是又极为有意义的研究工作。 |
| English Abstract: | Software cost estimation techniques abound, each with its own set of advantages and disadvantages, and no one stands out as the single best answer. Combining estimating is an appealing approach. Instead of trying to choose the single “best” technique, one frames the problem by asking which techniques would help to improve accuracy, assuming that each has something to contribute.
In this paper, we firstly introduce the systematic combining idea into the field of software cost estimating, we also introduce the current popular software cost estimating methods in detail and analyze their characteristics respectively, which is the premise of combining different methods to improve accuracy. Meanwhile, we also introduce the method of choosing complementary single methods for combining.
Then we estimate software effort using Optimal Linear Combining （OLC） method with a case study based on a real-life data set. The result indicates that combining different techniques can significantly improve the accuracy and consistency of software cost estimation by making full use of information provided by all components, even the much “worse” one.
We also discuss the factors which would influence the OLC’s effectiveness and extend this by doing an empirical data sensitivity analysis. This analysis helps us to find outliers that would have bad effect on the OLC’s effectiveness and the validation experiment also proves its effectiveness.
From our research in combining estimation for software cost, we prove this idea of estimation’s effectiveness and its help to improve the estimation’s accuracy and consistency. Combining forecasting has become one of the mainstreams in the field of forecasting, however, there is nearly blank in software cost estimation. We are eager to call for more attention to this challenging but meaningful research field. |
| Language: | 中文
|
| Content Type: | 学位论文
|
| URI: | http://ir.iscas.ac.cn/handle/311060/6742
|
| Appears in Collections: | 中科院软件所
|
| File Name/ File Size |
Content Type |
Version |
Access |
License |
|
|---|
| 10001_200528015029031李奇_paper.doc（2249KB） | -- | -- | 限制开放 | -- | 联系获取全文 |
|
| Recommended Citation: | |
李奇. 软件成本最优线性组合估算方法与应用研究[D]. 软件研究所. 中国科学院软件研究所. 2008-05-29.
|
|
|
