add emotion analysis, release version 1.2

mesolitica · Jan 6, 2019 · 01ba1da · 01ba1da
1 parent 06e0be1
commit 01ba1da
Show file tree

Hide file tree

Showing 133 changed files with 18,580 additions and 7,146 deletions.
diff --git a/README.rst b/README.rst
@@ -33,6 +33,8 @@ GPU version
 Features
 --------
 
+-  **Emotion Analysis**, from BERT, Fast-Text, Dynamic-Memory Network, Sparse-Chars,
+   Attention to build deep emotion analysis models.
 -  **Entities Recognition**, using latest state-of-art CRF deep learning
    models to do Naming Entity Recognition.
 -  **Language Detection**, using Multinomial, SGD, XGB, Fast-text N-grams deep learning to distinguish Malay, English, and Indonesian.
@@ -41,12 +43,12 @@ Features
 -  Num2Word
 -  **Part-of-Speech Recognition**, using latest state-of-art CRF deep
    learning models to do POS Recognition.
--  **Sentiment Analysis**, from BERT, Fast-Text, Dynamic-Memory Network,
+-  **Sentiment Analysis**, from BERT, Fast-Text, Dynamic-Memory Network, Sparse-Chars,
    Attention to build deep sentiment analysis models.
 -  **Spell Correction**, using local Malaysia NLP researches to
    auto-correct any bahasa words.
 -  Stemmer
--  **Subjectivity Analysis**, from BERT, Fast-Text, Dynamic-Memory Network,
+-  **Subjectivity Analysis**, from BERT, Fast-Text, Dynamic-Memory Network, Sparse-Chars,
    Attention to build deep subjectivity analysis models.
 -  **Summarization**, using skip-thought state-of-art to give precise
    summarization.

diff --git a/accuracy/emotion-accuracy.png b/accuracy/emotion-accuracy.png
diff --git a/accuracy/emotion-template.js b/accuracy/emotion-template.js
@@ -0,0 +1,28 @@
+option = {
+    xAxis: {
+        type: 'category',
+        axisLabel: {
+            interval: 0,
+            rotate: 30
+        },
+        data: ['bahdanau','BERT','bidirectional','entity-network',
+        'fast-text','fast-text-char','hierarchical','luong',
+        'multinomial','xgb']
+    },
+    yAxis: {
+        type: 'value',
+        min:0.73,
+        max:0.81
+    },
+    backgroundColor:'rgb(252,252,252)',
+    series: [{
+        data: [0.79,0.77,0.80,0.76,0.77,0.75,0.80,0.79,0.75,0.79],
+        type: 'bar',
+        label: {
+                normal: {
+                    show: true,
+                    position: 'top'
+                }
+            },
+    }]
+};
diff --git a/accuracy/models-accuracy.ipynb b/accuracy/models-accuracy.ipynb
diff --git a/accuracy/models-accuracy.rst b/accuracy/models-accuracy.rst
@@ -432,6 +432,18 @@ Fast-text
 
    avg / total       0.73      0.73      0.73      2856
 
+Fast-text-char
+^^^^^^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+      negative       0.71      0.64      0.67      1303
+      positive       0.72      0.78      0.75      1553
+
+   avg / total       0.71      0.71      0.71      2856
+
 Hierarchical
 ^^^^^^^^^^^^
 
@@ -499,7 +511,7 @@ Labels are,
 
 
 
-.. image:: models-accuracy_files/models-accuracy_33_0.png
+.. image:: models-accuracy_files/models-accuracy_34_0.png
    :width: 500px
 
 
@@ -628,7 +640,7 @@ sessions stored in
 
 
 
-.. image:: models-accuracy_files/models-accuracy_42_0.png
+.. image:: models-accuracy_files/models-accuracy_43_0.png
    :width: 500px
 
 
@@ -702,6 +714,18 @@ Fast-text
       macro avg       0.89      0.89      0.89      1993
    weighted avg       0.89      0.89      0.89      1993
 
+Fast-text-char
+^^^^^^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+      negative       0.88      0.88      0.88      1002
+      positive       0.88      0.87      0.88       991
+
+   avg / total       0.88      0.88      0.88      1993
+
 Hierarchical
 ^^^^^^^^^^^^
 
@@ -758,3 +782,180 @@ XGB
       micro avg       0.85      0.85      0.85      1993
       macro avg       0.85      0.85      0.85      1993
    weighted avg       0.85      0.85      0.85      1993
+
+Emotion Analysis
+----------------
+
+Trained on 80% of dataset, tested on 20% of dataset. All training
+sessions stored in
+`session/emotion <https://github.com/huseinzol05/Malaya/tree/master/session/emotion>`__
+
+.. code:: ipython3
+
+    display(Image('emotion-accuracy.png', width=500))
+
+
+
+.. image:: models-accuracy_files/models-accuracy_55_0.png
+   :width: 500px
+
+
+Bahdanau
+^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.80      0.80      0.80      3827
+          fear       0.77      0.78      0.78      3760
+           joy       0.81      0.78      0.80      3958
+          love       0.82      0.86      0.84      3099
+       sadness       0.73      0.76      0.74      3119
+      surprise       0.79      0.74      0.77      1940
+
+   avg / total       0.79      0.79      0.79     19703
+
+BERT
+^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.73      0.83      0.78      3747
+          fear       0.70      0.84      0.77      3789
+           joy       0.74      0.80      0.77      3929
+          love       0.82      0.76      0.79      3081
+       sadness       0.82      0.60      0.69      3168
+      surprise       0.85      0.63      0.72      1989
+
+   avg / total       0.77      0.76      0.76     19703
+
+Bidirectional
+^^^^^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.81      0.80      0.81      3726
+          fear       0.77      0.78      0.77      3806
+           joy       0.83      0.81      0.82      3975
+          love       0.86      0.83      0.85      2992
+       sadness       0.75      0.78      0.77      3293
+      surprise       0.77      0.79      0.78      1911
+
+   avg / total       0.80      0.80      0.80     19703
+
+Entity-network
+^^^^^^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.82      0.72      0.77      3717
+          fear       0.72      0.77      0.75      3743
+           joy       0.77      0.74      0.76      4050
+          love       0.81      0.81      0.81      2992
+       sadness       0.71      0.74      0.72      3274
+      surprise       0.72      0.80      0.76      1927
+
+   avg / total       0.76      0.76      0.76     19703
+
+Fast-text
+^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.82      0.75      0.78      3754
+          fear       0.71      0.81      0.75      3837
+           joy       0.76      0.79      0.78      3844
+          love       0.83      0.83      0.83      3065
+       sadness       0.75      0.75      0.75      3241
+      surprise       0.79      0.64      0.71      1962
+
+   avg / total       0.77      0.77      0.77     19703
+
+Fast-text-char
+^^^^^^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.79      0.75      0.77      3803
+          fear       0.73      0.73      0.73      3784
+           joy       0.71      0.77      0.74      3872
+          love       0.81      0.80      0.80      3052
+       sadness       0.72      0.70      0.71      3205
+      surprise       0.73      0.70      0.72      1987
+
+   avg / total       0.75      0.74      0.75     19703
+
+Hierarchical
+^^^^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.81      0.79      0.80      3786
+          fear       0.78      0.79      0.78      3754
+           joy       0.81      0.82      0.82      3886
+          love       0.85      0.84      0.85      3022
+       sadness       0.76      0.80      0.78      3300
+      surprise       0.81      0.75      0.78      1955
+
+   avg / total       0.80      0.80      0.80     19703
+
+Luong
+^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.80      0.79      0.80      3774
+          fear       0.78      0.75      0.77      3759
+           joy       0.79      0.80      0.79      3944
+          love       0.83      0.84      0.84      3033
+       sadness       0.75      0.75      0.75      3272
+      surprise       0.76      0.80      0.78      1921
+
+   avg / total       0.79      0.79      0.79     19703
+
+Multinomial
+^^^^^^^^^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.72      0.82      0.77      3833
+          fear       0.68      0.80      0.74      3802
+           joy       0.68      0.84      0.75      3924
+          love       0.85      0.71      0.78      2981
+       sadness       0.81      0.67      0.73      3189
+      surprise       0.80      0.36      0.50      1974
+
+   avg / total       0.75      0.73      0.73     19703
+
+XGB
+^^^
+
+.. code:: text
+
+                precision    recall  f1-score   support
+
+         anger       0.80      0.80      0.80      3769
+          fear       0.79      0.76      0.78      3808
+           joy       0.79      0.81      0.80      3913
+          love       0.84      0.85      0.84      2998
+       sadness       0.76      0.75      0.76      3250
+      surprise       0.77      0.77      0.77      1965
+
+   avg / total       0.79      0.79      0.79     19703
diff --git a/accuracy/models-accuracy_files/models-accuracy_14_0.png b/accuracy/models-accuracy_files/models-accuracy_14_0.png
diff --git a/accuracy/models-accuracy_files/models-accuracy_21_0.png b/accuracy/models-accuracy_files/models-accuracy_21_0.png
diff --git a/accuracy/models-accuracy_files/models-accuracy_22_0.png b/accuracy/models-accuracy_files/models-accuracy_22_0.png
diff --git a/...s-accuracy_files/models-accuracy_32_0.png → ...s-accuracy_files/models-accuracy_34_0.png b/...s-accuracy_files/models-accuracy_32_0.png → ...s-accuracy_files/models-accuracy_34_0.png
diff --git a/accuracy/models-accuracy_files/models-accuracy_41_0.png b/accuracy/models-accuracy_files/models-accuracy_41_0.png
diff --git a/accuracy/models-accuracy_files/models-accuracy_42_0.png b/accuracy/models-accuracy_files/models-accuracy_42_0.png
diff --git a/accuracy/models-accuracy_files/models-accuracy_43_0.png b/accuracy/models-accuracy_files/models-accuracy_43_0.png
diff --git a/accuracy/models-accuracy_files/models-accuracy_55_0.png b/accuracy/models-accuracy_files/models-accuracy_55_0.png
diff --git a/accuracy/models-accuracy_files/models-accuracy_9_0.png b/accuracy/models-accuracy_files/models-accuracy_9_0.png
diff --git a/accuracy/sentiment-accuracy.png b/accuracy/sentiment-accuracy.png
diff --git a/accuracy/sentiment-template.js b/accuracy/sentiment-template.js
@@ -0,0 +1,28 @@
+option = {
+    xAxis: {
+        type: 'category',
+        axisLabel: {
+            interval: 0,
+            rotate: 30
+        },
+        data: ['bahdanau','BERT','bidirectional','entity-network',
+        'fast-text','fast-text-char','hierarchical','luong',
+        'multinomial','xgb']
+    },
+    yAxis: {
+        type: 'value',
+        min:0.65,
+        max:0.75
+    },
+    backgroundColor:'rgb(252,252,252)',
+    series: [{
+        data: [0.67,0.69,0.67,0.71,0.73,0.71,0.67,0.66,0.73,0.69],
+        type: 'bar',
+        label: {
+                normal: {
+                    show: true,
+                    position: 'top'
+                }
+            },
+    }]
+};
diff --git a/accuracy/subjectivity-accuracy.png b/accuracy/subjectivity-accuracy.png
diff --git a/accuracy/subjectivity-template.js b/accuracy/subjectivity-template.js
@@ -0,0 +1,28 @@
+option = {
+    xAxis: {
+        type: 'category',
+        axisLabel: {
+            interval: 0,
+            rotate: 30
+        },
+        data: ['bahdanau','BERT','bidirectional','entity-network',
+        'fast-text','fast-text-char','hierarchical','luong',
+        'multinomial','xgb']
+    },
+    yAxis: {
+        type: 'value',
+        min:0.81,
+        max:0.9
+    },
+    backgroundColor:'rgb(252,252,252)',
+    series: [{
+        data: [0.83,0.84,0.85,0.88,0.89,0.88,0.84,0.82,0.89,0.85],
+        type: 'bar',
+        label: {
+                normal: {
+                    show: true,
+                    position: 'top'
+                }
+            },
+    }]
+};
diff --git a/docs/<MagicMock name='mock.home()' id='4471680920'>/Malaya/stop-word-kerulnet b/docs/<MagicMock name='mock.home()' id='4471680920'>/Malaya/stop-word-kerulnet
diff --git a/docs/<MagicMock name='mock.home()' id='4471680920'>/Malaya/version b/docs/<MagicMock name='mock.home()' id='4471680920'>/Malaya/version
@@ -0,0 +1 @@
+1.0
diff --git a/docs/<MagicMock name='mock.home()' id='4519654400'>/Malaya/stop-word-kerulnet b/docs/<MagicMock name='mock.home()' id='4519654400'>/Malaya/stop-word-kerulnet
diff --git a/docs/<MagicMock name='mock.home()' id='4519654400'>/Malaya/version b/docs/<MagicMock name='mock.home()' id='4519654400'>/Malaya/version
@@ -0,0 +1 @@
+1.2
diff --git a/docs/Accuracy.rst b/docs/Accuracy.rst
@@ -1,4 +1,4 @@
-Models accuracy
-====================
+Models Accuracy
+============================
 
 .. include:: models-accuracy.rst
diff --git a/docs/Contribution.rst b/docs/Contribution.rst
@@ -4,10 +4,14 @@ Contribution
 Rules
 -----
 
-Malaya has some standard code like PEP8 if want to add, we use Black Mamba to prettify our code.
+1. Malaya has some standard code like PEP8 if want to add, we use Black Mamba to prettify our code.
 
 You can install it from here, https://github.com/mohtar/blackmamba.
 
+2. 100% Tensorflow, no Keras.
+
+3.  3.3 > Python < 3.7
+
 New code
 --------