Skip to first unread message
Birone L
Mar 25, 2014, 5:10:37 AM3/25/14
to wekamooc...@googlegroups.com
1) The grader only seems to accept answers if the capitalisation matches exactly; it's easy to get this wrong (eg if someone's switching between the Weka and browser windows).
2) It might be useful to specify in the question that the relevant classify option takes a numeric value (argument); it causes the classify tab to fail silently (& irrecoverably?) during "setting up". (A bit frustrating if you don't like referring to the answer prematurely.)
Anyway, thanks to Prof Witten & everyone involved for a great MOOC & a very nifty piece of software! :)
Ian Witten
Mar 26, 2014, 8:55:40 PM3/26/14
to wekamooc...@googlegroups.com
On 25/03/2014, at 10:10 pm, Birone L <biron...@gmail.com> wrote:
1) The grader only seems to accept answers if the capitalisation matches exactly; it's easy to get this wrong (eg if someone's switching between the Weka and browser windows).
Um … the regular expression that checks answers [is intended to] accept correct capitalisation, or all lower case. (It doesn’t accept all upper case.) If that’s not how it appears to you, please send me an example.
Actually, Google’s CourseBuilder, the platform we use for this MOOC, is pretty rudimentary in terms of what you can do with free-text answers (e.g. if you ask for a number, you can only check it with a regular expression (!); there’s no way of giving feedback for alternative not-quite-correct answers). Oh well.
2) It might be useful to specify in the question that the relevant classify option takes a numeric value (argument); it causes the classify tab to fail silently (& irrecoverably?) during "setting up". (A bit frustrating if you don't like referring to the answer prematurely.)
Good point. I have added a small note to the question that will appear when we next deploy the software, at the beginning of next week. I’ve also requested that the silent, catastrophic, failure be fixed in future versions of Weka.
cheers
ian
Anyway, thanks to Prof Witten & everyone involved for a great MOOC & a very nifty piece of software! :)
--
You received this message because you are subscribed to the Google Groups "WekaMOOC-general" group.
To unsubscribe from this group and stop receiving emails from it, send an email to wekamooc-gener...@googlegroups.com.
To post to this group, send email to wekamooc...@googlegroups.com.
Visit this group at http://groups.google.com/group/wekamooc-general.
To view this discussion on the web, visit https://groups.google.com/d/msgid/wekamooc-general/6ca1cd81-b2fb-487a-8eea-7090fae6d2a5%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Birone L
Mar 27, 2014, 5:56:27 AM3/27/14
to wekamooc...@googlegroups.com
If that’s not how it appears to you, please send me an example.
Ok, it looks like mangling the capitalisation in a creative ways is what throws it:
Actually, Google’s CourseBuilder, the platform we use for this MOOC, is pretty rudimentary in terms of what you can do with free-text answers (e.g. if you ask for a number, you can only check it with a regular expression (!);
Way to go, Google ;)
Good point. I have added a small note to the question that will appear when we next deploy the software, at the beginning of next week. I’ve also requested that the silent, catastrophic, failure be fixed in future versions of Weka.
Great, that really had me scratching my head!
DogCat
Apr 5, 2014, 11:11:34 PM4/5/14
to wekamooc...@googlegroups.com
Hi
After adding the extra attribute through AddID filter:
Just noticed after adding the extra attributeID to the prediction output, the prediction results are different? Or i cant find an instance # with the same prediction of 0.32 for the third instance. Is this supposed to happen?
Thanks
-----------Iris Setosa
Relation: iris-weka.filters.unsupervised.attribute.MakeIndicator-Clast-V1
Instances: 150
Attributes: 5
sepallength
sepalwidth
petallength
petalwidth
class
Test mode:10-fold cross-validation
=== Classifier model (full training set) ===
Linear Regression Model
class =
0.079 * sepallength +
0.228 * sepalwidth +
-0.2561 * petallength +
0.1382
Time taken to build model: 0 seconds
=== Predictions on test data ===
inst#, actual, predicted, error
1 0 -0.065 -0.065
2 0 0.015 0.015
3 0 0.32 0.32
4 1 0.827 -0.173
5 0 0.102 0.102
6 0 -0.276 -0.276
7 0 0.175 0.175
8 0 -0.154 -0.154
=== Run information ===
Scheme:weka.classifiers.functions.LinearRegression -S 0 -R 1.0E-8
Relation: iris-weka.filters.unsupervised.attribute.MakeIndicator-Clast-V1-weka.filters.unsupervised.attribute.AddID-Cfirst-NID
Instances: 150
Attributes: 6
ID
sepallength
sepalwidth
petallength
petalwidth
class
Test mode:10-fold cross-validation
=== Classifier model (full training set) ===
Linear Regression Model
class =
0.079 * sepallength +
0.228 * sepalwidth +
-0.2561 * petallength +
0.1382
Time taken to build model: 0 seconds
=== Predictions on test data ===
inst#, actual, predicted, error (ID)
1 0 -0.077 -0.077 (141)
2 0 0.009 0.009 (146)
3 0 0.293 0.293 (65)
4 1 0.805 -0.195 (4)
5 0 0.06 0.06 (71)
6 0 -0.281 -0.281 (106)
7 0 0.191 0.191 (82)
8 0 -0.159 -0.159 (133)
9 1 0.988 -0.012 (6)
10 0 -0.104 -0.104 (144)
11 0 0.04 0.04 (78)
12 1 1.047 0.047 (37)
13 1 0.893 -0.107 (7)
14 0 0.047 0.047 (127)
15 1 0.824 -0.176 (2)
1 0 0.288 0.288 (80)
2 0 0.117 0.117 (55)
3 1 0.843 -0.157 (38)
4 1 0.952 -0.048 (32)
5 1 0.865 -0.135 (48)
6 0 -0.071 -0.071 (117)
7 0 0.221 0.221 (76)
8 0 0.109 0.109 (67)
9 1 0.991 -0.009 (41)
10 0 -0.122 -0.122 (103)
11 0 -0.019 -0.019 (150)
12 1 1.056 0.056 (23)
13 1 0.839 -0.161 (46)
14 1 0.914 -0.086 (44)
15 1 0.961 -0.039 (29)
1 0 -0.067 -0.067 (126)
2 0 0.385 0.385 (99)
3 0 -0.077 -0.077 (84)
4 0 0.239 0.239 (72)
5 1 1.104 0.104 (17)
6 1 0.991 -0.009 (22)
7 0 0.178 0.178 (87)
8 0 0.148 0.148 (60)
9 0 0.237 0.237 (83)
10 0 0.201 0.201 (93)
11 0 -0.023 -0.023 (69)
12 0 0.172 0.172 (70)
13 0 0.024 0.024 (124)
14 0 0.074 0.074 (74)
15 1 0.967 -0.033 (50)
1 0 -0.461 -0.461 (119)
2 1 0.835 -0.165 (24)
3 0 -0.232 -0.232 (108)
4 0 0.215 0.215 (66)
5 0 0.064 0.064 (77)
6 1 1.014 0.014 (19)
7 0 0.065 0.065 (128)
8 0 -0.077 -0.077 (121)
9 0 0.135 0.135 (92)
10 1 0.808 -0.192 (26)
11 0 0.025 0.025 (54)
12 0 -0.115 -0.115 (102)
13 0 0.193 0.193 (100)
14 0 -0.057 -0.057 (107)
15 0 -0.043 -0.043 (122)
1 0 0.097 0.097 (139)
2 1 0.959 -0.041 (18)
3 0 -0.073 -0.073 (143)
4 1 0.865 -0.135 (3)
5 1 0.944 -0.056 (40)
6 1 1.201 0.201 (34)
7 0 0.07 0.07 (111)
8 1 0.824 -0.176 (30)
9 0 0.186 0.186 (52)
10 1 1.048 0.048 (49)
11 0 -0.038 -0.038 (145)
12 1 0.826 -0.174 (10)
13 0 0.029 0.029 (140)
14 1 0.745 -0.255 (9)
15 1 0.818 -0.182 (13)
1 1 1.006 0.006 (11)
2 1 0.839 -0.161 (25)
3 0 0.219 0.219 (62)
4 1 1.01 0.01 (47)
5 1 1.032 0.032 (20)
6 1 0.883 -0.117 (21)
7 0 -0.27 -0.27 (109)
8 0 0.116 0.116 (61)
9 0 0.153 0.153 (81)
10 0 0.241 0.241 (89)
11 0 -0.14 -0.14 (130)
12 0 -0.394 -0.394 (123)
13 0 0.239 0.239 (94)
14 0 -0.073 -0.073 (113)
15 0 -0.09 -0.09 (147)
1 1 1.002 0.002 (5)
2 0 -0.239 -0.239 (135)
3 0 -0.054 -0.054 (110)
4 0 0.103 0.103 (63)
5 0 -0.076 -0.076 (138)
6 0 -0.124 -0.124 (118)
7 0 -0.156 -0.156 (104)
8 0 0.112 0.112 (95)
9 1 1.214 0.214 (16)
10 0 0.138 0.138 (53)
11 0 -0.201 -0.201 (131)
12 0 0.147 0.147 (59)
13 1 0.973 -0.027 (36)
14 0 -0.181 -0.181 (101)
15 0 0.045 0.045 (88)
1 1 0.607 -0.393 (42)
2 1 0.93 -0.07 (8)
3 0 0.188 0.188 (98)
4 0 -0.011 -0.011 (116)
5 0 0.126 0.126 (56)
6 0 -0.155 -0.155 (129)
7 1 0.961 -0.039 (28)
8 0 0.204 0.204 (97)
9 0 0.112 0.112 (64)
10 0 -0.165 -0.165 (120)
11 0 -0.01 -0.01 (137)
12 0 0.199 0.199 (68)
13 0 0.035 0.035 (149)
14 0 -0.138 -0.138 (105)
15 0 0.271 0.271 (86)
1 0 -0.002 -0.002 (148)
2 0 0.225 0.225 (58)
3 0 0.221 0.221 (96)
4 1 0.886 -0.114 (43)
5 0 0.033 0.033 (132)
6 1 1.13 0.13 (33)
7 0 0.128 0.128 (79)
8 1 0.939 -0.061 (45)
9 1 0.834 -0.166 (39)
10 0 0.199 0.199 (57)
11 0 0.048 0.048 (142)
12 0 0.117 0.117 (85)
13 1 1.198 0.198 (15)
14 1 0.82 -0.18 (31)
15 0 -0.027 -0.027 (125)
1 1 0.884 -0.116 (27)
2 1 0.888 -0.112 (12)
3 0 -0.011 -0.011 (134)
4 1 0.853 -0.147 (35)
5 0 0.108 0.108 (90)
6 1 0.869 -0.131 (14)
7 0 -0.132 -0.132 (115)
8 0 -0.154 -0.154 (114)
9 0 0.233 0.233 (51)
10 0 0.209 0.209 (75)
11 0 0.058 0.058 (91)
12 1 0.977 -0.023 (1)
13 0 -0.108 -0.108 (112)
14 0 -0.042 -0.042 (73)
15 0 -0.145 -0.145 (136)
Tyler Neill
Feb 24, 2017, 10:20:42 AM2/24/17
to WekaMOOC-general
For the answer to this question, I think it's best to refer to the similar ones by Shao-hui Lee and by RJ Wilderness.
-Tyler
Reply all
Reply to author
Forward
0 new messages