feat(eg): feature selection + filter method

Author: Ashwin Hegde

Feature_Selection.ipynb

@@ -0,0 +1,85 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Feature selection\n",
+    "\n",
+    "What is feature selection?\n",
+    "\n",
+    "* Feature or variable selection is the process of selecting a subset of relevant features from a total features available in a dataset to build ML algorithms.\n",
+    "\n",
+    "Why should we select features?\n",
+    "\n",
+    "* Its easy to understand the output which uses 10 variables as compare to 100 variables. Thus, simple models are easier to interpret.\n",
+    "* Shorter training times.\n",
+    "* Reduce the risk of data errors during model use.\n",
+    "* Reduce the variable redundancy. Exclude co-related variables.\n",
+    "* Bad learning behaviour in high dimensional space."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Feature selection methods\n",
+    "\n",
+    "1. Filter methods\n",
+    "2. Wrapper methods\n",
+    "3. Embedded methods"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Wrapper methods\n",
+    "\n",
+    "* Use predictive ML models to score the feature subset.\n",
+    "* Train a new model on each feature subset.\n",
+    "* Tend to be very computationally expensive.\n",
+    "* They may not produce the best feature combination for a different ML model.\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Embedded methods\n",
+    "\n",
+    "* Perform feature selection as part of the model construction process.\n",
+    "* Consider the interaction between features and models.\n",
+    "* They are less computationally expensive than wrapper methods, because they fit the ML model only once."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": []
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.6.0"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}