Skip to content
This repository was archived by the owner on Mar 8, 2021. It is now read-only.

Commit e17ff1f

Browse files
fbkarsdorpfbkarsdorp
committed
remove answers
1 parent 5cf1e7d commit e17ff1f

1 file changed

Lines changed: 9 additions & 52 deletions

File tree

‎Chapter 9 - Learning from Examples.ipynb‎

Lines changed: 9 additions & 52 deletions
Original file line numberDiff line numberDiff line change
@@ -1,7 +1,7 @@
11
{
22
"metadata": {
33
"name": "",
4-
"signature": "sha256:0d06f7c0dcbe740f33bc07e6b2f3b75bd99464160080eeb0dc89c4b381ba494a"
4+
"signature": "sha256:cfdf2e2f87626e61a47356de271a32958896c29bce6ec4af4456a82dc25b3c9b"
55
},
66
"nbformat": 3,
77
"nbformat_minor": 0,
@@ -208,7 +208,6 @@
208208
" def fit(self, X, y):\n",
209209
" \"Find that label in the training data that occurs most often.\"\n",
210210
" # insert your code here\n",
211-
" self.most_popular = max(set(y), key=y.count)\n",
212211
" \n",
213212
" def predict(self, x):\n",
214213
" \"Always predict that `x`'s label is equal to the most popular one.\"\n",
@@ -498,10 +497,9 @@
498497
" \"\"\"Fit or train the naive bayes classifier. X must be an\n",
499498
" iterable of examples where each example is an iterable as\n",
500499
" well.\"\"\"\n",
501-
" self.C = Counter(y) # insert your code here (class counts)\n",
502-
" self.N = defaultdict(Counter) # insert your code here (feature counts per class)\n",
503-
" for x, y_x in zip(X, y):\n",
504-
" self.N[y_x] += Counter(x)\n",
500+
" self.C = # insert your code here (class counts)\n",
501+
" self.N = # insert your code here (feature counts per class)\n",
502+
" # add the feature counts per class here\n",
505503
" \n",
506504
" def predict(self, x):\n",
507505
" \"\"\"Predict the outcome for example x. Choose the most\n",
@@ -575,15 +573,12 @@
575573
"\n",
576574
" def prior(self, y):\n",
577575
" \"\"\"Return the prior probability of class y.\"\"\"\n",
578-
" # insert your code here\n",
579-
" # pass\n",
580-
" return self.C[y] / sum(self.C.values())\n",
576+
" pass\n",
581577
" \n",
582578
" def probability(self, x, y):\n",
583579
" \"\"\"Apply Laplace Smoothing to give a probability\n",
584580
" estimate of feature x given y.\"\"\"\n",
585581
" # insert your code here\n",
586-
" return (self.N[y][x] + 1.0) / (sum(self.N[y].values()) + len(self.N))\n",
587582
"\n",
588583
"# these tests should return True if your code is correct\n",
589584
"nb = NaiveBayesLearner()\n",
@@ -720,11 +715,6 @@
720715
" \"\"\"Predict the outcome for example x. Choose the most\n",
721716
" likely outcome out of all possible outcomes.\"\"\"\n",
722717
" # insert your code here\n",
723-
" x_counts = Counter(x)\n",
724-
" def score(y):\n",
725-
" return log(self.prior(y)) + sum(log(self.probability(x_i, y) * c) \n",
726-
" for x_i, c in x_counts.items())\n",
727-
" return max(self.N.keys(), key=score)\n",
728718
"\n",
729719
"# these tests should return True if your code is correct\n",
730720
"nb = NaiveBayesLearner()\n",
@@ -863,10 +853,7 @@
863853
" iterable = tuple(iterable)\n",
864854
" ngrams = []\n",
865855
" # insert your code here\n",
866-
" for n in range(min_n, min(max_n + 1, len(iterable) + 1)):\n",
867-
" for i in range(len(iterable) - n + 1):\n",
868-
" ngrams.append(iterable[i: i + n])\n",
869-
" return ngrams # leave this in students version\n",
856+
" return ngrams\n",
870857
"\n",
871858
"# these tests should return True if your code is correct\n",
872859
"print(ngrams(\"humanities\", ngram_range=(2, 3)) == ['hu', 'um', 'ma', 'an', 'ni', 'it', \n",
@@ -985,13 +972,9 @@
985972
" \"\"\"Fit or train the Authorship Learner. X must be an\n",
986973
" iterable of examples where each example is an iterable as\n",
987974
" well.\"\"\"\n",
988-
" # insert your code here\n",
989975
" self.C = Counter(y)\n",
990976
" self.N = defaultdict(Counter)\n",
991-
" for x, y_x in zip(X, y):\n",
992-
" self.N[y_x] += Counter(x)\n",
993-
" for y_x, counts in self.N.items():\n",
994-
" self.N[y_x] = Counter(dict(counts.most_common(self.n_most_frequent)))\n",
977+
" # insert your code here\n",
995978
" \n",
996979
"# these tests should return True if your code is correct\n",
997980
"X = [('a', 'a', 'a', 'b', 'b', 'c'), ('a', 'c', 'b', 'b'), ('b', 'b', 'b', 'c', 'c', 'a')]\n",
@@ -1075,7 +1058,6 @@
10751058
" \"\"\"Return the accuracy, defined as the proportion of\n",
10761059
" correctly predicted labels.\"\"\"\n",
10771060
" # insert your code here\n",
1078-
" return sum(pred == true for pred, true in zip(pred_labels, true_labels)) / len(true_labels)\n",
10791061
"\n",
10801062
"# these tests should return True if your code is correct\n",
10811063
"predictions = [1, 2, 0, 0, 1]\n",
@@ -1206,13 +1188,6 @@
12061188
" if matrix is None:\n",
12071189
" matrix = error_statistics(pred_labels, true_labels)\n",
12081190
" # insert your code here\n",
1209-
" scores = {}\n",
1210-
" for label in set(true_labels):\n",
1211-
" try:\n",
1212-
" scores[label] = matrix[label]['tp'] / (matrix[label]['tp'] + matrix[label]['fn'])\n",
1213-
" except ZeroDivisionError:\n",
1214-
" scores[label] = 0.0\n",
1215-
" return scores if not average else mean(scores.values())\n",
12161191
" \n",
12171192
"# these tests should return True if your code is correct\n",
12181193
"predictions = [1, 2, 0, 0, 1]\n",
@@ -1238,15 +1213,6 @@
12381213
" if matrix is None:\n",
12391214
" matrix = error_statistics(pred_labels, true_labels)\n",
12401215
" # insert your code here\n",
1241-
" p = precision(pred_labels, true_labels, average=False, matrix=matrix)\n",
1242-
" r = recall(pred_labels, true_labels, average=False, matrix=matrix)\n",
1243-
" scores = {}\n",
1244-
" for label in set(true_labels):\n",
1245-
" try:\n",
1246-
" scores[label] = 2 * (p[label] * r[label]) / (p[label] + r[label])\n",
1247-
" except ZeroDivisionError:\n",
1248-
" scores[label] = 0.0\n",
1249-
" return scores if not average else mean(scores.values())\n",
12501216
"\n",
12511217
"# these tests should return True if your code is correct\n",
12521218
"predictions = [1, 2, 0, 0, 1]\n",
@@ -1451,11 +1417,7 @@
14511417
"collapsed": false,
14521418
"input": [
14531419
"scores = {}\n",
1454-
"# insert your code here\n",
1455-
"for sample in range(100, 5000, 500):\n",
1456-
" documents = [make_document(f, sample=sample) for f in glob('data/british-novels/*.txt')]\n",
1457-
" authors, titles, texts = zip(*documents)\n",
1458-
" scores[sample] = cross_validate(AuthorshipLearner(), texts, authors, k=None, score_fn=f_score)"
1420+
"# insert your code here"
14591421
],
14601422
"language": "python",
14611423
"metadata": {},
@@ -1493,12 +1455,7 @@
14931455
"collapsed": false,
14941456
"input": [
14951457
"scores = {}\n",
1496-
"# insert your code here\n",
1497-
"for n_most_frequent in range(50, 500, 100):\n",
1498-
" documents = [make_document(f) for f in glob('data/british-novels/*.txt')]\n",
1499-
" authors, titles, texts = zip(*documents)\n",
1500-
" scores[n_most_frequent] = cross_validate(AuthorshipLearner(n_most_frequent=n_most_frequent), \n",
1501-
" texts, authors, k=None, score_fn=f_score)"
1458+
"# insert your code here"
15021459
],
15031460
"language": "python",
15041461
"metadata": {},

0 commit comments

Comments
 (0)