Fixed typos in chapter 13 notebook

Rethinking/Chp_13.ipynb

@@ -406,7 +406,7 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "For complicate model it is always good to do more checks:"
+    "For complex models it is always good to do more checks:"
    ]
   },
   {
@@ -854,10 +854,10 @@
     "    corr = tt.diag(sigma_ab**-1).dot(cov.dot(tt.diag(sigma_ab**-1)))\n",
     "    r = pm.Deterministic('Rho', corr[np.triu_indices(2, k=1)])\n",
     "    \n",
-    "    mu = pm.MvNormal('ab_cafe', mu=tt.stack([a, bm]), chol=chol, shape=(Ndept, 2))\n",
+    "    mu = pm.MvNormal('ab_dept', mu=tt.stack([a, bm]), chol=chol, shape=(Ndept, 2))\n",
     "    \n",
     "    a_dept = pm.Deterministic('a_dept', mu[:, 0])\n",
-    "    a_dept = pm.Deterministic('bm_dept', mu[:, 1])\n",
+    "    bm_dept = pm.Deterministic('bm_dept', mu[:, 1])\n",
     "    \n",
     "    p = pm.math.invlogit(mu[Dept_id, 0] + mu[Dept_id, 1] * d_ad['male'])\n",
     "    admit = pm.Binomial('admit', p=p, n=d_ad.applications, observed=d_ad.admit)\n",
@@ -1491,7 +1491,7 @@
    "metadata": {},
    "source": [
     "#### Code 13.28\n",
-    "Actually in model m_13_6NC I am already using the Cholesky decomposition of the covariance matrix. If we want to strictly follow the parameterization of m13.6NC and m13.6nc1 as in the book, we will model the Rho using pm.LKJCorr and do multiplication it with pm.HalfCauchy."
+    "Actually in model m_13_6NC I am already using the Cholesky decomposition of the covariance matrix. If we want to strictly follow the parameterization of m13.6NC and m13.6nc1 as in the book, we will model Rho using pm.LKJCorr and multiply it with pm.HalfCauchy."
    ]
   },
   {
@@ -2096,6 +2096,7 @@
     }
    ],
    "source": [
+    "# Summary of centered model\n",
     "pm.summary(trace_13_7, varnames=['g', 'a', 'bp', 'etasq', 'rhosq']).round(2)"
    ]
   },
@@ -2303,6 +2304,7 @@
     }
    ],
    "source": [
+    "# Summary of non-centered model\n",
     "pm.summary(trace_13_7_, varnames=['g', 'a', 'bp', 'etasq', 'rhosq']).round(2)"
    ]
   },
@@ -2348,7 +2350,7 @@
     "\n",
     "ax.set_ylim(0, 1)\n",
     "ax.set_xlabel('distance')\n",
-    "ax.set_ylabel('correlation');"
+    "ax.set_ylabel('covariance');"
    ]
   },
   {
@@ -2656,21 +2658,23 @@
     }
    ],
    "source": [
+    "# compute posterior median relationship, ignoring distance\n",
     "Nsamp, Nbin = 1000, 30\n",
     "log_pop_seq = np.linspace(6, 14, Nbin)\n",
     "a_post = trace_13_7.get_values(varname='a', combine=True)[:, None]\n",
     "bp_post = trace_13_7.get_values(varname='bp', combine=True)[:, None]\n",
     "lambda_post = np.exp(a_post + bp_post*log_pop_seq)\n",
     "\n",
-    "# plot raw data\n",
     "_, axes = plt.subplots(1, 1, figsize=(5, 5))\n",
-    "axes.plot(log_pop_seq, np.median(lambda_post, axis=0), '--', color='k')\n",
-    "\n",
     "alpha = .2\n",
+    "\n",
+    "# display posterior predictions\n",
+    "axes.plot(log_pop_seq, np.median(lambda_post, axis=0), '--', color='k')\n",
     "mu_hpd = pm.hpd(lambda_post, alpha=alpha)\n",
     "axes.fill_between(log_pop_seq,\n",
     "                  mu_hpd[:,0], mu_hpd[:,1], alpha=alpha*.5+.1, color='k')\n",
     "\n",
+    "# plot raw data and labels\n",
     "axes.scatter(dk['logpop'], dk['total_tools'], psize)\n",
     "labels = dk['culture'].values\n",
     "for i, itext in enumerate(labels):\n",
@@ -2726,9 +2730,9 @@
  "metadata": {
   "anaconda-cloud": {},
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": "stat-rethink-pymc3",
    "language": "python",
-   "name": "python3"
+   "name": "stat-rethink-pymc3"
   },
   "language_info": {
    "codemirror_mode": {
@@ -2740,7 +2744,7 @@
    "name": "python",
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
-   "version": "3.6.3"
+   "version": "3.7.3"
   }
  },
  "nbformat": 4,