Simplify log ratios in Trask notebooks

The original calculations of log ratios for words with negative sentiment are unnecessarily complex. Simplify them and add some explanation.
sebastianstock · May 19, 2017 · 6f855ae · 6f855ae
1 parent d1541a8
commit 6f855ae
Show file tree

Hide file tree

Showing 2 changed files with 2,949 additions and 2,809 deletions.
diff --git a/sentiment-network/Sentiment_Classification_Projects.ipynb b/sentiment-network/Sentiment_Classification_Projects.ipynb
@@ -77,6 +77,7 @@
    "cell_type": "code",
    "execution_count": null,
    "metadata": {
+    "collapsed": true,
     "nbpresent": {
      "id": "eba2b193-0419-431e-8db9-60f34dd3fe83"
     }
@@ -105,7 +106,9 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "len(reviews)"
@@ -115,6 +118,7 @@
    "cell_type": "code",
    "execution_count": null,
    "metadata": {
+    "collapsed": true,
     "nbpresent": {
      "id": "bb95574b-21a0-4213-ae50-34363cf4f87f"
     }
@@ -128,6 +132,7 @@
    "cell_type": "code",
    "execution_count": null,
    "metadata": {
+    "collapsed": true,
     "nbpresent": {
      "id": "e0408810-c424-4ed4-afb9-1735e9ddbd0a"
     }
@@ -148,6 +153,7 @@
    "cell_type": "code",
    "execution_count": null,
    "metadata": {
+    "collapsed": true,
     "nbpresent": {
      "id": "e67a709f-234f-4493-bae6-4fb192141ee0"
     }
@@ -196,7 +202,9 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "# Create three Counter objects to store positive, negative and total counts\n",
@@ -318,11 +326,9 @@
     "\n",
     "To fix these issues, we'll convert all of our ratios to new values using logarithms.\n",
     "\n",
-    "**TODO:** Go through all the ratios you calculated and convert their values using the following formulas:\n",
-    "> * For any postive words, convert the ratio using `np.log(ratio)`\n",
-    "> * For any negative words, convert the ratio using `-np.log(1/(ratio + 0.01))`\n",
+    "**TODO:** Go through all the ratios you calculated and convert them to logarithms. (i.e. use `np.log(ratio)`)\n",
     "\n",
-    "That second equation may look strange, but what it's doing is dividing one by a very small number, which will produce a larger positive number. Then, it takes the `log` of that, which produces numbers similar to the ones for the postive words. Finally, we negate the values by adding that minus sign up front. In the end, extremely positive and extremely negative words will have positive-to-negative ratios with similar magnitudes but oppositite signs."
+    "In the end, extremely positive and extremely negative words will have positive-to-negative ratios with similar magnitudes but opposite signs."
    ]
   },
   {
@@ -415,7 +421,9 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "from IPython.display import Image\n",
@@ -428,7 +436,9 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "review = \"The movie was excellent\"\n",
@@ -448,7 +458,9 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "# TODO: Create set named \"vocab\" containing all of the words from all of the reviews\n",
@@ -484,7 +496,9 @@
   {
    "cell_type": "code",
    "execution_count": 1,
-   "metadata": {},
+   "metadata": {
+    "collapsed": false
+   },
    "outputs": [
     {
      "data": {
@@ -543,7 +557,9 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "from IPython.display import Image\n",
@@ -1082,7 +1098,9 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "metadata": {},
+   "metadata": {
+    "collapsed": true
+   },
    "outputs": [],
    "source": [
     "from IPython.display import Image\n",
@@ -1854,7 +1872,7 @@
  "metadata": {
   "anaconda-cloud": {},
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": "Python [default]",
    "language": "python",
    "name": "python3"
   },
@@ -1868,7 +1886,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.5.3"
+   "version": "3.5.2"
   }
  },
  "nbformat": 4,