572 project

1. prepare json files (already in data.tar.gz)
2. download project
   • run extract-repo-url.py, it will generate repo.txt and star.txt
   • run download.sh against repo.txt
3. run ground-all.sh against the folder containing project zip files. This will output ground raw output file.
4. run parse-ground.py against the raw ground output file will output csv data file.

Result table

SVM with radial

Number of discretize category	linear	polynomial	radial basis	sigmoid
2
3
4
5
6
7
8
9
10
cate	C	java	javascript
2
3
4

star range	precision
<500
500-1000
1000-1500
1500-2000
>2000

fold validation	precision
3
5
7
10

model	precision
lm
SVM + linear
SVM + polynormial
SVM + radial basis
SVM + sigmoid
Decision Tree
language	precision
C
java
javascript
shell
ruby

Tables

category	accuracy
2	0.83
3	0.71
4	0.61
5	0.54
6	0.47
7	0.42
8	0.38
9	0.35
10	0.32

category	c.csv	php.csv	java.csv	javascript.csv	shell.csv	ruby.csv	python.csv
2	0.92	0.93	0.84	0.56	0.95	0.88	0.83
3	0.85	0.87	0.7	0.39	0.92	0.78	0.74
4	0.79	0.82	0.61	0.31	0.87	0.69	0.63
5	0.76	0.76	0.52	0.22	0.82	0.63	0.56
6	0.72	0.71	0.47	0.17	0.8	0.58	0.51
7	0.67	0.67	0.42	0.15	0.77	0.53	0.46
8	0.63	0.63	0.36	0.15	0.74	0.49	0.42
9	0.6	0.59	0.33	0.11	0.71	0.46	0.39
10	0.57	0.57	0.3	0.1	0.69	0.44	0.35

category	linear	polynomial	radial	sigmoid
2	0.82	0.82	0.83	0.8
3	0.71	0.69	0.71	0.67
4	0.61	0.58	0.61	0.58
5	0.54	0.49	0.54	0.5
6	0.47	0.43	0.47	0.45
7	0.42	0.36	0.43	0.4
8	0.38	0.33	0.38	0.36
9	0.35	0.29	0.35	0.33
10	0.32	0.26	0.32	0.29