Merge pull request #87 from AnFreTh/main

Release 0.1.6

README.md

@@ -102,6 +102,10 @@ STREAM offers a variety of neural as well as non-neural topic models and we are
         <td><a href="https://www.jmlr.org/papers/volume3/blei03a/blei03a.pdf?ref=http://githubhelp.com">LDA</a></td>
         <td>Latent Dirichlet Allocation</td>
       </tr>
+      <tr>
+        <td><a href="https://www.nature.com/articles/44565">NMF</a></td>
+        <td>Non-negative Matrix Factorization</td>
+      </tr>
       <tr>
         <td><a href="https://arxiv.org/abs/2004.14914">WordCluTM</a></td>
         <td>Tired of topic models?</td>

docs/api/models/models.rst

@@ -2,11 +2,6 @@ Models
 ========
 
 
-.. autoclass:: stream_topic.models.BERTopicTM
-    :members:
-
-.. autoclass:: stream_topic.models.CBC
-    :members:
 
 .. autoclass:: stream_topic.models.CEDC
     :members:
@@ -40,3 +35,12 @@ Models
 
 .. autoclass:: stream_topic.models.CTMNeg
     :members:
+
+.. autoclass:: stream_topic.models.NMFTM
+    :members:
+
+.. autoclass:: stream_topic.models.BERTopicTM
+    :members:
+
+.. autoclass:: stream_topic.models.CBC
+    :members:

docs/landingpage.md

@@ -51,6 +51,10 @@ STREAM offers a variety of neural as well as non-neural topic models and we are
         <td><a href="https://www.jmlr.org/papers/volume3/blei03a/blei03a.pdf?ref=http://githubhelp.com">LDA</a></td>
         <td>Latent Dirichlet Allocation</td>
       </tr>
+      <tr>
+        <td><a href="https://www.nature.com/articles/44565">NMF</a></td>
+        <td>Non-negative Matrix Factorization</td>
+      </tr>
       <tr>
         <td><a href="https://arxiv.org/abs/2004.14914">WordCluTM</a></td>
         <td>Tired of topic models?</td>

stream_topic/__version__.py

@@ -1,4 +1,4 @@
 """Version information."""
 
 # The following line *must* be the last in the module, exactly as formatted:
-__version__ = "0.1.5"
+__version__ = "0.1.6"

stream_topic/models/__init__.py

@@ -12,6 +12,7 @@
 from .neurallda import NeuralLDA
 from .ctmneg import CTMNeg
 from .tntm import TNTM
+from .nmf import NMFTM
 
 __all__ = [
     "BERTopicTM",
@@ -28,4 +29,5 @@
     "NeuralLDA",
     "CTMNeg",
     "TNTM",
+    "NMFTM",
 ]

stream_topic/models/abstract_helper_models/base.py

@@ -343,8 +343,9 @@ def optimize_hyperparameters(
         assert criterion in [
             "aic",
             "bic",
+            "recon",
             "custom",
-        ], "Criterion must be either 'aic', 'bic', or 'custom'."
+        ], "Criterion must be either 'aic', 'bic', 'recon' or 'custom'."
         if criterion == "custom":
             assert (
                 custom_metric is not None
@@ -363,12 +364,14 @@ def objective(trial):
             self.fit(dataset)
 
             # Calculate the score based on the criterion
-            if criterion in ["aic", "bic"]:
+            if criterion in ["aic", "bic", "recon"]:
 
                 if criterion == "aic":
                     score = self.calculate_aic(n_topics=self.hparams["n_topics"])
-                else:
+                elif criterion == "bic":
                     score = self.calculate_bic(n_topics=self.hparams["n_topics"])
+                elif criterion == "recon":
+                    score = self.reconstruction_loss()
             else:
                 # Compute the custom metric score
                 topics = self.get_topics()