Wygładzanie Alpha

run.py

@@ -0,0 +1,26 @@
+import pandas as pd
+from collections import defaultdict
+from utils import read_csv
+from train import train_model, predict_data
+
+
+def main():
+    data = read_csv("train/in.tsv.xz")
+    train_words = read_csv("train/expected.tsv")
+
+    train_data = data[[6, 7]]
+    train_data = pd.concat([train_data, train_words], axis=1)
+    train_data["607"] = train_data[6] + train_data[0] + train_data[7]
+
+    model = defaultdict(lambda: defaultdict(lambda: 0))
+    alpha = 0.00002
+    vocab = set()
+
+    train_model(train_data, model,vocab,alpha)
+
+    predict_data("dev-0/in.tsv.xz", "dev-0/out.tsv", model)
+    predict_data("test-A/in.tsv.xz", "test-A/out.tsv", model)
+
+
+if __name__ == "__main__":
+    main()

train.py

@@ -0,0 +1,57 @@
+from collections import Counter
+from nltk import bigrams, word_tokenize
+from utils import read_csv, ENCODING, clean_text
+
+DEFAULT_PREDICTION = "the:0.2 be:0.2 to:0.2 of:0.1 and:0.1 a:0.1 :0.1"
+
+def train_model(data, model,vocab,alpha):
+    for _, row in data.iterrows():
+        words = word_tokenize(clean_text(row["607"]))
+        for w1, w2 in bigrams(words, pad_left=True, pad_right=True):
+            if w1 and w2:
+                model[w2][w1] += 1
+                vocab.add(w2)
+                vocab.add(w1)
+
+    for w2 in model:
+        total_count = float(sum(model[w2].values()))
+        denominator = total_count + alpha * len(vocab)
+        for w1 in model[w2]:
+            nominator = model[w2][w1] + alpha
+            model[w2][w1] /= nominator / denominator
+
+
+def predict_data(read_path, save_path, model):
+    data = read_csv(read_path)
+
+    with open(save_path, "w", encoding=ENCODING) as f:
+        for _, row in data.iterrows():
+            words = word_tokenize(clean_text(row[7]))
+            if len(words) < 3:
+                prediction = DEFAULT_PREDICTION
+            else:
+                prediction = predict(words[0], model)
+            f.write(prediction + "\n")
+
+
+def predict(word, model):
+    predictions = dict(model[word])
+    most_common = dict(Counter(predictions).most_common(6))
+
+    total_prob = 0.0
+    str_prediction = ""
+
+    for word, prob in most_common.items():
+        total_prob += prob
+        str_prediction += f"{word}:{prob} "
+
+    if total_prob == 0.0:
+        return DEFAULT_PREDICTION
+
+    rem_prob = 1 - total_prob
+    if rem_prob < 0.01:
+        rem_prob = 0.01
+
+    str_prediction += f":{rem_prob}"
+
+    return str_prediction

utils.py

@@ -0,0 +1,33 @@
+import pandas as pd
+import regex as re
+from csv import QUOTE_NONE
+
+ENCODING = "utf-8"
+
+REP = re.compile(r"[{}\[\]\&%^$*#\(\)@\t\n0123456789]+")
+REM = re.compile(r"'s|[\-]\\n|\-\\n|\p{P}")
+
+def read_csv(fname):
+    return pd.read_csv(
+        fname,
+        sep="\t",
+        on_bad_lines='skip',
+        header=None,
+        quoting=QUOTE_NONE,
+        encoding=ENCODING
+    )
+
+def clean_text(text):
+    res = str(text).lower().strip()
+    res = res.replace("’", "'")
+    res = REM.sub("", res)
+    res = REP.sub(" ", res)
+    res = res.replace("'s", " is")
+    res = res.replace("'ll", " will")
+    res = res.replace("won't", "will not")
+    res = res.replace("isn't", "is not")
+    res = res.replace("aren't", "are not")
+    res = res.replace("'ve'", "have")
+    return res.replace("'m", " am")
+
+