242 lines
8.9 KiB
Markdown
242 lines
8.9 KiB
Markdown
file:: [icse2023-paper223_1666614236200_0.pdf](../assets/icse2023-paper223_1666614236200_0.pdf)
|
||
file-path:: ../assets/icse2023-paper223_1666614236200_0.pdf
|
||
|
||
- technical debt prediction
|
||
ls-type:: annotation
|
||
hl-page:: 1
|
||
hl-color:: yellow
|
||
id:: 6356cb9b-96f6-43e9-aff8-2a144ff92bd7
|
||
- consensus clustering approac
|
||
ls-type:: annotation
|
||
hl-page:: 1
|
||
hl-color:: purple
|
||
id:: 6356cbc0-068c-400c-b1cd-e2f3cf1333aa
|
||
hl-stamp:: 1666632648322
|
||
- It is advocated that the proposed clustering environment could be useful for facilitating the decision making process for business investors and open-source community with the help of the Gartner’s hype cycle
|
||
ls-type:: annotation
|
||
hl-page:: 1
|
||
hl-color:: yellow
|
||
id:: 6356cbee-4989-4235-80dc-00e293e40bfe
|
||
- n this work, we aim to overcome the limitations (such as limited validity) of using a single clustering method with our main objectives along with key aspects of originality being the following:
|
||
ls-type:: annotation
|
||
hl-page:: 1
|
||
hl-color:: yellow
|
||
id:: 6356cce6-4135-4399-bf28-03118e86e9ae
|
||
- onsensus clusterin
|
||
ls-type:: annotation
|
||
hl-page:: 1
|
||
hl-color:: purple
|
||
id:: 6356cd24-872f-41d9-9ecc-dafb1860b57b
|
||
hl-stamp:: 1666633011154
|
||
- multiple cluster validity indices
|
||
ls-type:: annotation
|
||
hl-page:: 1
|
||
hl-color:: purple
|
||
id:: 6356cd3a-d252-42a7-bb63-a8c7603f2886
|
||
- The relevance of this study can be mostly attributed to studying the software ecosystems and identifying important indicators about development activity in software projects
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: yellow
|
||
id:: 6356cd98-c4af-44f7-9b3f-0ac16e70e4a2
|
||
- prototype” vectors
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: purple
|
||
id:: 6356cdd5-9133-44a6-89f3-8c55aaa20ae1
|
||
- analyze the results
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356ce1a-9ec6-4a49-9391-a217d12baf56
|
||
hl-stamp:: 1666633245706
|
||
- details significant parts of the methodology such as clustering with application of validity indices, consensus clustering, and aggregation of the results
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356ce2b-6c3d-4184-80fb-d370b96c1a8e
|
||
- web libraries and frameworks
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356cf26-2514-416b-a73c-08abe6018aad
|
||
- non-web libraries and frameworks
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356cf31-580a-45aa-aeb3-7e76a875a96c
|
||
- software tools
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356cf50-79ae-4bda-b29f-a7d9dfe0d678
|
||
- Commit-related event
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356cfe8-55f0-40a6-9066-3f816a9237df
|
||
- Issue-related even
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356cff2-464f-4716-a566-62c345bc7ee6
|
||
- Pull request-related even
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: green
|
||
id:: 6356cffa-43e0-42b2-a11e-a004d5744905
|
||
- ] ,
|
||
ls-type:: annotation
|
||
hl-page:: 2
|
||
hl-color:: red
|
||
id:: 6356d017-7e72-4e11-956f-35aaeb5eda5e
|
||
- ,
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: red
|
||
id:: 6356d032-b692-475a-b46c-3577fa01e677
|
||
- we try to identify the metrics that would characterize software repositories qualitatively and signal about potential anomalies
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: green
|
||
id:: 6356d03c-785d-4fca-adf3-93fdc6881e0b
|
||
- “value simple and easily understandable metrics over complex ones
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: green
|
||
id:: 6356d087-1588-4870-823e-24378961c127
|
||
- To address this, we conduct an automatic clustering of the repositories based on a set of GitHub metrics instead of relying on the simple “extreme outlier” definition of anomaly, and then try to identify useful metrics that indicate something unusual in the resulting clusters of repositories.
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: yellow
|
||
id:: 6356d0ca-0058-46ec-8e80-c7ddd3269913
|
||
- 2.1.
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: red
|
||
id:: 6356d0ec-71b7-4970-8f96-c0f5157c57fb
|
||
- The limitation of such clustering is that it produces too element-specific, not descriptive, and small clusters.
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: green
|
||
id:: 6356d12d-2c44-4132-9390-bbcf9ece62ca
|
||
- two cluster
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: yellow
|
||
id:: 6356d170-ce29-41c0-8190-f09d18bc968c
|
||
hl-stamp:: 1666634101493
|
||
- Tsoukalas et al. (2021) [10] divided 27 software projects from the technical debt dataset [11] into six clusters of similar projects with respect to their technical debt aspects using Kmeans algorithm and built specific technical debt forecasting models for each cluster using regression algorithms.
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: green
|
||
id:: 6356d26a-d805-47a2-b4a7-9197fee85841
|
||
- The results showed that the prediction errors tend to be statistically significantly lower in within-cluster technical debt forecasting than in cross-cluster forecastin
|
||
ls-type:: annotation
|
||
hl-page:: 3
|
||
hl-color:: green
|
||
id:: 6356d4b0-eb9c-475e-8850-b54692ea0bef
|
||
- In this work, we try to overcome the limitation of cluster validity by using multiple clustering techniques, cluster validity indices, and consensus clusterin
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: yellow
|
||
id:: 6356d572-7838-40b2-9ebb-20965c0d258c
|
||
- onsensus clustering to analyze software repository metric
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: yellow
|
||
id:: 6356d5dd-19ba-49bc-b071-6caa9796a72c
|
||
- ize or popularity based clustering
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: green
|
||
id:: 6356d72b-4f96-412b-b62c-99b47ed1280e
|
||
- data collection and preparation
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: green
|
||
id:: 6356d7a9-213a-49ca-9be7-ec4dec1adaa4
|
||
- redefine the fixed number of consensus clusters
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: green
|
||
id:: 6356d7ba-ee81-4511-b1da-2637eef1f5d7
|
||
- We normalize each metric to the [0,1] range.
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: green
|
||
id:: 6356d7ec-cdc7-4fae-ae05-b915bdbb8a32
|
||
- m clustering algorithms with different values of parameters that we tune.
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: green
|
||
id:: 6356d7f6-93be-4ce5-9583-27e270f6ff4a
|
||
- three cluster validity indices
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: yellow
|
||
id:: 6356d806-f610-4275-8e96-1f5862b94d31
|
||
- winner”
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: yellow
|
||
id:: 6356d870-8683-46aa-8db3-3a471d7804d9
|
||
- we get a collection of c sets of prototype vectors
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: yellow
|
||
id:: 6356d8af-e6d2-4a51-8426-d9c284b7b72c
|
||
- The final step of our methodology is aggregation of the results. We match the calculated cluster prototypes to compare them and the clusters they represent across P sets of clustered repositories. For that, we consider the minimum-weight matching problem in P -partite graph across all c · P prototypes (for details refer to Section III-D), and, as a result, get c sets of P matched prototypes. We calculate the discrepancy in every such set of prototype vectors and, using a similarity measure, calculate the discrepancy value d for each of the c sets, which shows us how different the cluster prototypes are across the P sets of data for each of the c clusters.
|
||
ls-type:: annotation
|
||
hl-page:: 4
|
||
hl-color:: yellow
|
||
id:: 6356d962-696b-481d-b08c-3098461da7f7
|
||
- if two out of three indices showed the same solution, and the third one showed a new solution, we would consider the two unique solution
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: green
|
||
id:: 6356d9e3-5da3-4fc8-bd09-91a14d5d5e37
|
||
- if all three showed the same solution or different solutions, we would consider only one or three unique solutions correspondingly
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: green
|
||
id:: 6356d9f2-adb2-49ec-8718-0396690a8ff9
|
||
- Silhouette, Calinski-Harabasz, and DaviesBouldin
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: green
|
||
id:: 6356dab9-6876-4531-8cb4-1709cb6f1160
|
||
- mean nearest-cluster distanc
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: yellow
|
||
id:: 6356e087-e141-4bc5-8e1a-4fa042d7e72a
|
||
- ntra-cluster distance
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: yellow
|
||
id:: 6356e08b-5698-4395-9a70-46d162b6423d
|
||
- silhouette coefficient
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: yellow
|
||
id:: 6356e303-65ce-48a3-abe4-4f86b2978aa0
|
||
- Cohesion is estimated based on the distances from the repositories in a cluster to its cluster centroid and separation is based on the distance of the cluster centroids from the global centroid
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: yellow
|
||
id:: 6356e43d-fe11-4b65-a710-80a46abb4a00
|
||
- cluster ensemble
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: yellow
|
||
id:: 6356e4e1-85db-4df1-9014-cc8cdbe7dbd8
|
||
- quare matrix S
|
||
ls-type:: annotation
|
||
hl-page:: 5
|
||
hl-color:: green
|
||
id:: 6356e5af-06b1-4057-9dc4-962699459775
|
||
- Meta-clustering algorithm is based on clustering clusters of software repositorie
|
||
ls-type:: annotation
|
||
hl-page:: 6
|
||
hl-color:: yellow
|
||
id:: 6356e5f8-1a4a-442c-9fa4-fc3ec528a7ee |