exercise 2:sentiment-analysis-rf-with-bert-encodings.ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"!pip install hazm","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import numpy as np\nimport pandas as pd\nimport hazm\n\nfrom sklearn.ensemble import RandomForestClassifier\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import classification_report\n\nfrom transformers import BertTokenizer","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:03.920630Z","iopub.execute_input":"2023-11-02T19:46:03.921481Z","iopub.status.idle":"2023-11-02T19:46:17.797684Z","shell.execute_reply.started":"2023-11-02T19:46:03.921420Z","shell.execute_reply":"2023-11-02T19:46:17.796861Z"},"trusted":true},"execution_count":1,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/scipy/__init__.py:146: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.24.3\n  warnings.warn(f\"A NumPy version >={np_minversion} and <{np_maxversion}\"\n","output_type":"stream"}]},{"cell_type":"code","source":"data = pd.read_csv('/kaggle/input/snappfood-persian-sentiment-analysis/Snappfood - Sentiment Analysis.csv', on_bad_lines='skip' , delimiter='\\t')\ndata.head()","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:17.799222Z","iopub.execute_input":"2023-11-02T19:46:17.799724Z","iopub.status.idle":"2023-11-02T19:46:18.236176Z","shell.execute_reply.started":"2023-11-02T19:46:17.799695Z","shell.execute_reply":"2023-11-02T19:46:18.235254Z"},"trusted":true},"execution_count":2,"outputs":[{"execution_count":2,"output_type":"execute_result","data":{"text/plain":"  Unnamed: 0                                            comment  label  \\\n0        NaN    واقعا حیف وقت که بنویسم سرویس دهیتون شده افتضاح    SAD   \n1        NaN  قرار بود ۱ ساعته برسه ولی نیم ساعت زودتر از مو...  HAPPY   \n2        NaN  قیمت این مدل اصلا با کیفیتش سازگاری نداره، فقط...    SAD   \n3        NaN  عالللی بود همه چه درست و به اندازه و کیفیت خوب...  HAPPY   \n4        NaN                      شیرینی وانیلی فقط یک مدل بود.  HAPPY   \n\n   label_id  \n0       1.0  \n1       0.0  \n2       1.0  \n3       0.0  \n4       0.0  ","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>Unnamed: 0</th>\n      <th>comment</th>\n      <th>label</th>\n      <th>label_id</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>NaN</td>\n      <td>واقعا حیف وقت که بنویسم سرویس دهیتون شده افتضاح</td>\n      <td>SAD</td>\n      <td>1.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>NaN</td>\n      <td>قرار بود ۱ ساعته برسه ولی نیم ساعت زودتر از مو...</td>\n      <td>HAPPY</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>NaN</td>\n      <td>قیمت این مدل اصلا با کیفیتش سازگاری نداره، فقط...</td>\n      <td>SAD</td>\n      <td>1.0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>NaN</td>\n      <td>عالللی بود همه چه درست و به اندازه و کیفیت خوب...</td>\n      <td>HAPPY</td>\n      <td>0.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>NaN</td>\n      <td>شیرینی وانیلی فقط یک مدل بود.</td>\n      <td>HAPPY</td>\n      <td>0.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"cell_type":"code","source":"data.dropna(inplace=True, subset=['label', 'label_id'])\ndata.drop_duplicates(keep='first', inplace=True)\ndata['label_id'] = data['label_id'].astype(int)","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:18.237262Z","iopub.execute_input":"2023-11-02T19:46:18.237569Z","iopub.status.idle":"2023-11-02T19:46:18.361094Z","shell.execute_reply.started":"2023-11-02T19:46:18.237542Z","shell.execute_reply":"2023-11-02T19:46:18.359845Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"code","source":"data['comment_len'] = data['comment'].apply(lambda t: len(hazm.word_tokenize(t)))","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:18.363678Z","iopub.execute_input":"2023-11-02T19:46:18.364012Z","iopub.status.idle":"2023-11-02T19:46:21.046129Z","shell.execute_reply.started":"2023-11-02T19:46:18.363984Z","shell.execute_reply":"2023-11-02T19:46:21.045238Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"data['comment_len'] = data['comment_len'].apply(lambda len_t: len_t if len_t <= 100 else None)\ndata = data.dropna(subset=['comment_len'])\ndata = data.reset_index(drop=True)","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:21.047485Z","iopub.execute_input":"2023-11-02T19:46:21.047883Z","iopub.status.idle":"2023-11-02T19:46:21.101931Z","shell.execute_reply.started":"2023-11-02T19:46:21.047828Z","shell.execute_reply":"2023-11-02T19:46:21.100881Z"},"trusted":true},"execution_count":5,"outputs":[]},{"cell_type":"code","source":"normalizer = hazm.Normalizer()\ndata['normalized_comment'] = data['comment'].apply(lambda c: normalizer.normalize(c.strip()))","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:21.103209Z","iopub.execute_input":"2023-11-02T19:46:21.103510Z","iopub.status.idle":"2023-11-02T19:46:49.386616Z","shell.execute_reply.started":"2023-11-02T19:46:21.103481Z","shell.execute_reply":"2023-11-02T19:46:49.385781Z"},"trusted":true},"execution_count":6,"outputs":[]},{"cell_type":"code","source":"data = data[['normalized_comment','label_id']]\ndata.columns = ['comment', 'label_id']\ndata.head()","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:49.387698Z","iopub.execute_input":"2023-11-02T19:46:49.387981Z","iopub.status.idle":"2023-11-02T19:46:49.402341Z","shell.execute_reply.started":"2023-11-02T19:46:49.387956Z","shell.execute_reply":"2023-11-02T19:46:49.401359Z"},"trusted":true},"execution_count":7,"outputs":[{"execution_count":7,"output_type":"execute_result","data":{"text/plain":"                                             comment  label_id\n0    واقعا حیف وقت که بنویسم سرویس دهیتون شده افتضاح         1\n1  قرار بود ۱ ساعته برسه ولی نیم ساعت زودتر از مو...         0\n2  قیمت این مدل اصلا با کیفیتش سازگاری نداره، فقط...         1\n3  عالی بود همه چه درست و به اندازه و کیفیت خوب، ...         0\n4                      شیرینی وانیلی فقط یک مدل بود.         0","text/html":"<div>\n<style scoped>\n    .dataframe tbody tr th:only-of-type {\n        vertical-align: middle;\n    }\n\n    .dataframe tbody tr th {\n        vertical-align: top;\n    }\n\n    .dataframe thead th {\n        text-align: right;\n    }\n</style>\n<table border=\"1\" class=\"dataframe\">\n  <thead>\n    <tr style=\"text-align: right;\">\n      <th></th>\n      <th>comment</th>\n      <th>label_id</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>واقعا حیف وقت که بنویسم سرویس دهیتون شده افتضاح</td>\n      <td>1</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>قرار بود ۱ ساعته برسه ولی نیم ساعت زودتر از مو...</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>قیمت این مدل اصلا با کیفیتش سازگاری نداره، فقط...</td>\n      <td>1</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>عالی بود همه چه درست و به اندازه و کیفیت خوب، ...</td>\n      <td>0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>شیرینی وانیلی فقط یک مدل بود.</td>\n      <td>0</td>\n    </tr>\n  </tbody>\n</table>\n</div>"},"metadata":{}}]},{"cell_type":"code","source":"MODEL_PATH = 'HooshvareLab/bert-fa-base-uncased'\ntokenizer = BertTokenizer.from_pretrained(MODEL_PATH)","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:49.403798Z","iopub.execute_input":"2023-11-02T19:46:49.404150Z","iopub.status.idle":"2023-11-02T19:46:50.575192Z","shell.execute_reply.started":"2023-11-02T19:46:49.404123Z","shell.execute_reply":"2023-11-02T19:46:50.574186Z"},"trusted":true},"execution_count":8,"outputs":[{"output_type":"display_data","data":{"text/plain":"Downloading (…)solve/main/vocab.txt:   0%|          | 0.00/1.20M [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"16b1db4e486f409d87e7f3056764d08f"}},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"Downloading (…)lve/main/config.json:   0%|          | 0.00/440 [00:00<?, ?B/s]","application/vnd.jupyter.widget-view+json":{"version_major":2,"version_minor":0,"model_id":"aa7e5c56b03f451c87abb37306501927"}},"metadata":{}}]},{"cell_type":"code","source":"input_ids = []\nmasks = []\nfor i in data.index:\n    encodings = tokenizer.encode_plus(\n        data['comment'][i],\n        max_length=32,\n        truncation=True,\n        add_special_tokens=True,\n        return_token_type_ids=True,\n        return_attention_mask=True,\n        padding='max_length',\n    )\n    input_ids.append(encodings.get(\"input_ids\"))\n    masks.append(encodings.get('attention_mask'))\n","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:46:50.576342Z","iopub.execute_input":"2023-11-02T19:46:50.576642Z","iopub.status.idle":"2023-11-02T19:47:43.671514Z","shell.execute_reply.started":"2023-11-02T19:46:50.576617Z","shell.execute_reply":"2023-11-02T19:47:43.670514Z"},"trusted":true},"execution_count":9,"outputs":[]},{"cell_type":"code","source":"print(len(data), len(input_ids), len(masks))","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:47:43.673991Z","iopub.execute_input":"2023-11-02T19:47:43.674293Z","iopub.status.idle":"2023-11-02T19:47:43.679383Z","shell.execute_reply.started":"2023-11-02T19:47:43.674268Z","shell.execute_reply":"2023-11-02T19:47:43.678511Z"},"trusted":true},"execution_count":10,"outputs":[{"name":"stdout","text":"69156 69156 69156\n","output_type":"stream"}]},{"cell_type":"code","source":"embeddings = np.multiply(input_ids, masks, dtype=int)\nembeddings[0]","metadata":{"execution":{"iopub.status.busy":"2023-11-02T19:47:43.680592Z","iopub.execute_input":"2023-11-02T19:47:43.680938Z","iopub.status.idle":"2023-11-02T19:47:44.174526Z","shell.execute_reply.started":"2023-11-02T19:47:43.680907Z","shell.execute_reply":"2023-11-02T19:47:44.173520Z"},"trusted":true},"execution_count":11,"outputs":[{"execution_count":11,"output_type":"execute_result","data":{"text/plain":"array([    2,  5668, 23317,  4244,  2800, 26283,  4875,  8113,  4957,\n        2871, 33691,     4,     0,     0,     0,     0,     0,     0,\n           0,     0,     0,     0,     0,     0,     0,     0,     0,\n           0,     0,     0,     0,     0])"},"metadata":{}}]},{"cell_type":"code","source":"y = (data['label_id'])\nX_train, X_test, y_train, y_test = train_test_split(embeddings, y, test_size=0.15)","metadata":{"execution":{"iopub.status.busy":"2023-11-02T20:16:58.518676Z","iopub.execute_input":"2023-11-02T20:16:58.519070Z","iopub.status.idle":"2023-11-02T20:16:58.551227Z","shell.execute_reply.started":"2023-11-02T20:16:58.519040Z","shell.execute_reply":"2023-11-02T20:16:58.550436Z"},"trusted":true},"execution_count":17,"outputs":[]},{"cell_type":"code","source":"randomFC = RandomForestClassifier(criterion='gini', n_estimators=1000)\n\nrandomFC.fit(X_train, y_train)\ny_pred = randomFC.predict(X_test)\nprint(classification_report(y_test, y_pred))","metadata":{"execution":{"iopub.status.busy":"2023-11-02T20:17:01.681318Z","iopub.execute_input":"2023-11-02T20:17:01.682035Z","iopub.status.idle":"2023-11-02T20:20:13.319535Z","shell.execute_reply.started":"2023-11-02T20:17:01.682001Z","shell.execute_reply":"2023-11-02T20:20:13.318412Z"},"trusted":true},"execution_count":18,"outputs":[{"name":"stdout","text":"              precision    recall  f1-score   support\n\n           0       0.74      0.60      0.66      5126\n           1       0.67      0.79      0.73      5248\n\n    accuracy                           0.70     10374\n   macro avg       0.70      0.70      0.69     10374\nweighted avg       0.70      0.70      0.69     10374\n\n","output_type":"stream"}]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}