11_datasets.ipynb: Add original tables

Add tables from FELINE, FERRARI and PReaCH papers as images, together
with description of columns.

Author: jnareb

11_datasets.ipynb

@@ -77,6 +77,22 @@
     "http://openprocedings.org/EDBT/2014/paper_166.pdf"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "![Table 1: Datasets](images/feline/FELINE_paper-table_1-datasets-with_caption.png)\n",
+    "\n",
+    "- _vertices_ - $|V|$, number of vertices / nodes in the graph\n",
+    "- _edges_ - $|E|$, number of edges in the graph\n",
+    "- _Cluster-coeff_ - clustering coefficient\n",
+    "- _Eff-diameter_ - effective diameter (or effective eccentricity) is an estimated size of the path in which 90% of all pairs of vertices connected are reachable from each other\n",
+    "- _roots_ - number of roots, nodes with no incoming edges, vertices with no predecessors (sources)\n",
+    "- _leafs_ - number of leafs, nodes with no outgoing edges (sinks)\n",
+    "\n",
+    "The FELINE paper authors used the SNAP software <snap.stanford.edu/snap/> to compute these values"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,
@@ -396,7 +412,20 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "\"FERRARI: Flexible and Efficient Reachability Range Assignment for Graph Indexing\" (2013)"
+    "\"FERRARI: Flexible and Efficient Reachability Range Assignment for Graph Indexing\" (2013)  \n",
+    "http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.365.2894&rep=rep1&type=pdf"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "<!-- ![Table 2. Datasets used](images/ferrari/FERRARI_arXiv-table_2-datasets_used-with_caption.png) -->\n",
+    "![Table 1. Datasets used](images/ferrari/FERRARI_paper-table_1-datasets_used-with_caption.png)\n",
+    "\n",
+    "- $|V|$ - number of vertices (nodes)\n",
+    "- $|E|$ - number of edges\n",
+    "- $|V_C|$, $|E_C|$ - number of, respectively, vertices and edges in the _condensed graph_, where maximial strongly connected components were collapsed into \"supernodes\", i.e. directed acyclic graph (DAG) created from given graph"
    ]
   },
   {
@@ -630,6 +659,18 @@
     "https://arxiv.org/abs/1404.4465"
    ]
   },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "![Table 1: Headers](images/preach/PReaCH_arXiv-table_1-datasets_used-header.png)  \n",
+    "![Table 1: Instances used for our experiments. _d_ is the maximal path length.](images/preach/PReaCH_arXiv-table_1-datasets_used-real_graphs-no_caption.png)\n",
+    "\n",
+    "- $m/n$ - edge density, number of edges divided by number of nodes (vertices)\n",
+    "- $d$ - length of the longest path, or maximal path length\n",
+    "- % pos - the fraction of positive queries in a random sample of 100000 queries"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": null,