Ainu/Transducer - Revision history

Dwurtel1 at 02:57, 27 February 2017

2017-02-27T02:57:02Z

2017-02-21T17:35:08Z

2017-02-21T16:51:10Z

2017-02-21T16:47:56Z

2017-02-18T20:34:01Z

‎Notes

2017-02-18T20:06:27Z

‎Notes

2017-02-18T19:48:55Z

‎Notes

2017-02-17T17:21:57Z

2017-02-17T17:20:50Z

‎Evaluation

2017-02-17T17:16:23Z

‎Evaluation

@@ Line 4: / Line 4: @@
 There are 79 stems in the transducer. 75/115 ain.yaml tests pass (40 fail). 2/20 commonwords.yaml tests pass (18 fail). Generation test has 77/164 ain.yaml tests pass (87 fail).
 === Notes ===
 Adding the .twoc prefix fix for getting rid of paths with odd numbers of feature tags fixed a lot. 77/133 ain.yaml tests pass, 56 fail.
 == Analyser ==

← Older revision		Revision as of 17:35, 21 February 2017
Line 4:		Line 4:

	There are 79 stems in the transducer. 75/115 ain.yaml tests pass (40 fail). 2/20 commonwords.yaml tests pass (18 fail). Generation test has 77/164 ain.yaml tests pass (87 fail).		There are 79 stems in the transducer. 75/115 ain.yaml tests pass (40 fail). 2/20 commonwords.yaml tests pass (18 fail). Generation test has 77/164 ain.yaml tests pass (87 fail).
		+
		+	=== Notes ===
		+
		+	Adding the .twoc prefix fix for getting rid of paths with odd numbers of feature tags fixed a lot. 77/133 ain.yaml tests pass, 56 fail.

	== Analyser ==		== Analyser ==

← Older revision		Revision as of 16:51, 21 February 2017
Line 1:		Line 1:
		+	== Generator ==

		+	=== Initial evaluation of morphological generation ===
		+
		+	There are 79 stems in the transducer. 75/115 ain.yaml tests pass (40 fail). 2/20 commonwords.yaml tests pass (18 fail). Generation test has 77/164 ain.yaml tests pass (87 fail).
		+
		+	== Analyser ==

	=== Evaluation ===		=== Evaluation ===

@@ Line 1: / Line 1: @@
 == Evaluation ==
 The most common unknown words originally were
@@ Line 69: / Line 71: @@
-Original coverage of these words is 11.19%. Now the coverage is 21.21%.
+Original coverage of these words is 11.19%. Now the coverage is 22.76%.
-There are 79 stems in the transducer. 58/115 ain.yaml tests pass (57 fail). 2/20 commonwords.yaml tests pass (18 fail).
+There are 79 stems in the transducer. 75/115 ain.yaml tests pass (40 fail). 2/20 commonwords.yaml tests pass (18 fail).
 Now the most common words are
@@ Line 141: / Line 143: @@
 |}
-== Notes ==
+=== Notes ===
 After adding "ネ<vbser> ↔ ネ" and "コ<cnjsub> ↔ コ" to the transducer, coverage went from 11.19% to 21.21%.
 The tests I am struggling with are for adjectivity-- I need a new grammar book to confirm that there are really so many differing adjective endings because the current book I am using as a reference was written over 100 years ago. I have not yet had time to add pronoun or noun affinity tests. The remaining failing tests are irregular cases at which I will have to look more closely.
-== Transducer ==
+=== Transducer ===
 Ainu transducer available on [https://github.swarthmore.edu/dwurtel1/ling073-ain my GitHub].
 [[Category:Sp17_Transducers]]

← Older revision		Revision as of 20:34, 18 February 2017
Line 147:		Line 147:

	It looks like the superlative イヨッタ will not be parsed properly as a prefix due to the 'small-tsu' character ッ. When I change the prefix to remove that character (i.e. イヨタ), the analysis of, for example, イヨタピリカ works (ピリカ<adj><sup>). Otherwise, with the prefix as it should be, the analysis of イヨッタピリカ is "^イ/イ$ヨ^ッタピリカ/ッタピリカ$".		It looks like the superlative イヨッタ will not be parsed properly as a prefix due to the 'small-tsu' character ッ. When I change the prefix to remove that character (i.e. イヨタ), the analysis of, for example, イヨタピリカ works (ピリカ<adj><sup>). Otherwise, with the prefix as it should be, the analysis of イヨッタピリカ is "^イ/イ$ヨ^ッタピリカ/ッタピリカ$".
		+
		+
		+	The diminutive tests are also slightly wonky, where ト゚カラポ will properly analyze as "^ト゚カラ<n><dim>$" but will fail the test claiming that analysis was not generated. It may have something to do with the diacritic? The test for オッカイポ fails because of the 'small-tsu' character which stops even オッカイ from being analyzed correctly.

	== Transducer ==		== Transducer ==

@@ Line 146: / Line 146: @@
 The tests I am struggling with are for adjectivity-- I need a new grammar book to confirm that there are really so many differing adjective endings because the current book I am using as a reference was written over 100 years ago. I have not yet had time to add pronoun or noun affinity tests. The remaining failing tests are irregular cases at which I will have to look more closely.
-It looks like the superlative イヨッタ will not be parsed properly as a prefix due to the 'small-tsu' character ッ. When I change the prefix to remove that character (i.e. イヨタ), the analysis works. Otherwise the analysis is "^イ/*イ$ヨ^ッタピリカ/*ッタピリカ$".
+It looks like the superlative イヨッタ will not be parsed properly as a prefix due to the 'small-tsu' character ッ. When I change the prefix to remove that character (i.e. イヨタ), the analysis of, for example, イヨタピリカ works (ピリカ<adj><sup>). Otherwise, with the prefix as it should be, the analysis of イヨッタピリカ is "^イ/*イ$ヨ^ッタピリカ/*ッタピリカ$".
 == Transducer ==

← Older revision		Revision as of 17:21, 17 February 2017
Line 149:		Line 149:
	== Transducer ==		== Transducer ==
	Ainu transducer available on [https://github.swarthmore.edu/dwurtel1/ling073-ain my GitHub].		Ainu transducer available on [https://github.swarthmore.edu/dwurtel1/ling073-ain my GitHub].
		+
		+	[[Category:Sp17_Transducers]]

← Older revision		Revision as of 17:16, 17 February 2017
Line 69:		Line 69:


−	Original coverage of these words is 11.19%.	+	Original coverage of these words is 11.19%. Now the coverage is 21.21%.
		+
		+	There are 79 stems in the transducer. 58/115 ain.yaml tests pass (57 fail). 2/20 commonwords.yaml tests pass (18 fail).

	== Notes ==		== Notes ==