{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.14","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":84896,"databundleVersionId":10305135,"sourceType":"competition"}],"dockerImageVersionId":30786,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Introduction","metadata":{"_uuid":"23568bf6-b483-4001-ad1d-86e3be74eed4","_cell_guid":"3a936ab5-7f05-401c-a532-f32d2a34477f","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"markdown","source":"Hi!\nThis is new playground competition and in this month we're having an insurance dataset. \n\nBefore diving deeper into code, let's describe a little topic in which we will work.\n\nFor start, a little meme which describes all this thing :)\n\n![](https://leadsurance.com/wp-content/uploads/2020/08/insurance-meme-6.jpg)\r\n","metadata":{}},{"cell_type":"markdown","source":"# Problem definition","metadata":{}},{"cell_type":"markdown","source":"## What is insurance?","metadata":{}},{"cell_type":"markdown","source":"
\n\n

Insurance is a means of protection from financial loss in which, in exchange for a fee, a party agrees to compensate another party in the event of a certain loss, damage, or injury. It is a form of risk management, primarily used to protect against the risk of a contingent or uncertain loss.

\n An entity which provides insurance is known as an insurer, insurance company, insurance carrier, or underwriter. A person or entity who buys insurance is known as a policyholder, while a person or entity covered under the policy is called an insured.\n\r\n

\n
","metadata":{}},{"cell_type":"markdown","source":"## How insurance Works\r
\n\n

\nWhen you buy a policy you make regular payments, known as premiums, to the insurer. If you make a claim your insurer will pay out for the loss that is covered under the policy.

\nIf you don’t make a claim, you won’t get your money back; instead it is pooled with the premiums of other policyholders who have taken out insurance with the same insurance company. If you make a claim the money comes from the pool of policyholders’ premiums.\n

\n
\n","metadata":{}},{"cell_type":"markdown","source":"## Insurance Policy Components\r\n","metadata":{}},{"cell_type":"markdown","source":"
\r\n\r\n\r\n\r\n
\r\n","metadata":{}},{"cell_type":"markdown","source":"# Import libraries","metadata":{"_uuid":"23e0c399-9c79-4374-8847-08df1b1784aa","_cell_guid":"096ad279-cbc5-49fd-b80c-9f34e221487e","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"import seaborn as sns\nimport matplotlib.pyplot as plt\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import mean_squared_log_error\nimport pandas as pd\nimport missingno as mnso\nimport numpy as np\nfrom sklearn.pipeline import Pipeline\nfrom sklearn.preprocessing import StandardScaler, OneHotEncoder\nfrom sklearn.compose import ColumnTransformer\nfrom sklearn.impute import SimpleImputer\nfrom sklearn.metrics import mean_squared_error\nfrom sklearn.ensemble import RandomForestRegressor\nfrom scipy.stats import boxcox\nfrom scipy.special import boxcox1p\n","metadata":{"_uuid":"90871205-85e8-45d8-97a6-490aee449832","_cell_guid":"f9eafd57-6931-4703-8bb0-eef4e35e86e7","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false},"execution":{"iopub.status.busy":"2024-12-18T20:16:54.758488Z","iopub.execute_input":"2024-12-18T20:16:54.75885Z","iopub.status.idle":"2024-12-18T20:16:55.856378Z","shell.execute_reply.started":"2024-12-18T20:16:54.758796Z","shell.execute_reply":"2024-12-18T20:16:55.855447Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Load data","metadata":{"_uuid":"d005054f-58e9-4091-b065-00c10dbfee18","_cell_guid":"229240ab-e333-4f0a-a562-7749e4add3aa","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"markdown","source":"After exploring discussions it seems reasonable to log transform our target variable to ensure that we optimizing directly the evaluation metric for this competition","metadata":{}},{"cell_type":"code","source":"X_train = pd.read_csv('/kaggle/input/playground-series-s4e12/train.csv')\nX_test = pd.read_csv('/kaggle/input/playground-series-s4e12/test.csv')\nsample_submission = pd.read_csv('/kaggle/input/playground-series-s4e12/sample_submission.csv')\n\nX_train = X_train.drop(['id'], axis=1)\nid_test = X_test.id\nX_test = X_test.drop('id', axis=1)\n# y = X_train['Premium Amount']\n# y_log = np.log1p(y)\n\npd.set_option('display.max_columns', None)","metadata":{"_uuid":"cbbb9373-1d55-4852-8548-09338a7de6d5","_cell_guid":"b811f64a-b2c4-4899-bdcf-9f948b67f048","trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:16:55.859006Z","iopub.execute_input":"2024-12-18T20:16:55.859578Z","iopub.status.idle":"2024-12-18T20:17:02.12822Z","shell.execute_reply.started":"2024-12-18T20:16:55.85953Z","shell.execute_reply":"2024-12-18T20:17:02.127291Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"X_train.columns","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:17:02.129447Z","iopub.execute_input":"2024-12-18T20:17:02.129801Z","iopub.status.idle":"2024-12-18T20:17:02.13745Z","shell.execute_reply.started":"2024-12-18T20:17:02.12976Z","shell.execute_reply":"2024-12-18T20:17:02.136607Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# EDA","metadata":{"_uuid":"55ee458e-7685-4061-a945-0fb6664d5fab","_cell_guid":"ff6fdb2b-3dd8-4946-b5c3-d8388ecd9432","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"X_train","metadata":{"_uuid":"145de48c-b9a5-4e07-bf4c-bb5884e16d5e","_cell_guid":"4f6a074b-f0e5-4eab-99c1-3392b5f7f2f7","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:02.138634Z","iopub.execute_input":"2024-12-18T20:17:02.139224Z","iopub.status.idle":"2024-12-18T20:17:02.1662Z","shell.execute_reply.started":"2024-12-18T20:17:02.139167Z","shell.execute_reply":"2024-12-18T20:17:02.165456Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"X_train.info()","metadata":{"_uuid":"d4cc7adb-0d6f-49dd-b3ad-f5b2c95d5202","_cell_guid":"5b9d679d-48e4-45f6-bbc3-5ccb5b1eca9a","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:02.167395Z","iopub.execute_input":"2024-12-18T20:17:02.168007Z","iopub.status.idle":"2024-12-18T20:17:02.718331Z","shell.execute_reply.started":"2024-12-18T20:17:02.167966Z","shell.execute_reply":"2024-12-18T20:17:02.717463Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"X_train['Number of Dependents'] = X_train['Number of Dependents'].astype('Int64')\nX_train['Vehicle Age'] = X_train['Vehicle Age'].astype('Int64')\nX_train['Insurance Duration'] = X_train['Insurance Duration'].astype('Int64')\nX_train['Previous Claims'] = X_train['Previous Claims'].astype('Int64')","metadata":{"_uuid":"9a16ab6e-08a3-4034-8baf-3d0ca5e511c8","_cell_guid":"f720fbc2-241b-4aa1-b974-687e33c3f0d8","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:02.719488Z","iopub.execute_input":"2024-12-18T20:17:02.71976Z","iopub.status.idle":"2024-12-18T20:17:03.380897Z","shell.execute_reply.started":"2024-12-18T20:17:02.719734Z","shell.execute_reply":"2024-12-18T20:17:03.379832Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"This features I decided to treat as discrete (maybe categorical), because they have fixed range of values","metadata":{}},{"cell_type":"markdown","source":"## Defining missing values","metadata":{"_uuid":"f5574e81-fad9-4a58-bb8d-fbaba0cd2b6f","_cell_guid":"bb69f33a-f5ff-46ab-8baf-ac2c5d5a66ab","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"mnso.bar(X_train)","metadata":{"_uuid":"ab35b427-5fb2-4c8d-8949-8ed2f3e85885","_cell_guid":"625d8c13-5048-486e-95c9-991df4c778fc","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:03.384765Z","iopub.execute_input":"2024-12-18T20:17:03.385613Z","iopub.status.idle":"2024-12-18T20:17:05.694922Z","shell.execute_reply.started":"2024-12-18T20:17:03.385568Z","shell.execute_reply":"2024-12-18T20:17:05.694045Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"It seems that we're having a smaller number of missing values than in previous competition.","metadata":{}},{"cell_type":"markdown","source":"## Defining feature's types","metadata":{"_uuid":"2a8f2644-611b-442d-a7ca-81511d599a47","_cell_guid":"5620dc3d-1d83-4602-9a35-0721e4a230e4","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"numerical_features_total = X_train.select_dtypes(exclude='object').columns\nnumerical_features_continuos = X_train.select_dtypes(exclude=['object', 'int']).columns\nnumerical_features_discrete = X_train.select_dtypes(exclude=['object', 'float']).columns\ncategorical_features = X_train.select_dtypes(include='object').columns","metadata":{"_uuid":"a3ed8ec3-5a89-4c2a-a95f-5a34e39dc2ec","_cell_guid":"7dcc0717-8747-4568-b7fe-f9017a3a7b9f","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:05.696063Z","iopub.execute_input":"2024-12-18T20:17:05.696325Z","iopub.status.idle":"2024-12-18T20:17:05.985119Z","shell.execute_reply.started":"2024-12-18T20:17:05.696299Z","shell.execute_reply":"2024-12-18T20:17:05.984285Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Heatmap for numerical features","metadata":{"_uuid":"0c30f7eb-f7dd-4765-8b3f-5259196f4226","_cell_guid":"6367a821-3cd8-407c-a58d-51cb62aa22e5","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"plt.figure(figsize=(10, 10))\nsns.heatmap(X_train[numerical_features_total].corr(), annot=True)","metadata":{"_uuid":"a0ac4a0c-298c-4769-bdae-9111875b7bec","_cell_guid":"2ca32ae5-0b17-4754-b4b4-b6e5c77734fc","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:05.986182Z","iopub.execute_input":"2024-12-18T20:17:05.98648Z","iopub.status.idle":"2024-12-18T20:17:06.901255Z","shell.execute_reply.started":"2024-12-18T20:17:05.986452Z","shell.execute_reply":"2024-12-18T20:17:06.900385Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"It seems that our features doesnt have a direct correlation with target variable (Premium Amount). In fact there's no strong correlation between every pair of features. Maybe situation will change after some feature engineering later.","metadata":{}},{"cell_type":"markdown","source":"## Numerical features summary","metadata":{"_uuid":"07c2007c-ab07-4e73-9351-ac3e23f42a50","_cell_guid":"2772a3ce-1bca-4d15-be2b-ed5c83708804","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"X_train[numerical_features_total].describe()","metadata":{"_uuid":"6685c723-cfff-4b6a-aace-eec17b8b834c","_cell_guid":"4179eaba-8d75-4baf-a0e1-d6bff698bf57","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:06.902503Z","iopub.execute_input":"2024-12-18T20:17:06.902851Z","iopub.status.idle":"2024-12-18T20:17:07.519228Z","shell.execute_reply.started":"2024-12-18T20:17:06.902813Z","shell.execute_reply":"2024-12-18T20:17:07.518273Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Vizualization of continuos features","metadata":{"_uuid":"64f3d234-c851-449a-a2fc-3e46b7c7abdd","_cell_guid":"bc2d3b2e-1646-4c3a-9411-b80f6ca02945","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"numerical_features_continuos","metadata":{"_uuid":"bd51bde0-fc63-491d-9861-f2e61b456a6c","_cell_guid":"03c15480-92d4-463d-bf72-8a5c216bbba5","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:07.520617Z","iopub.execute_input":"2024-12-18T20:17:07.521023Z","iopub.status.idle":"2024-12-18T20:17:07.528004Z","shell.execute_reply.started":"2024-12-18T20:17:07.520978Z","shell.execute_reply":"2024-12-18T20:17:07.527052Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"fig, axes = plt.subplots(2, 3, figsize=(15,15))\nfig.suptitle('Distributions of continuous features')\n\n\nsns.kdeplot(ax=axes[0,0], data=X_train, x='Credit Score')\naxes[0,0] = 'Credit Score Distribution'\nsns.kdeplot(ax=axes[0,1], data=X_train, x='Age')\naxes[0,1] = 'Age Distribution'\nsns.kdeplot(ax=axes[0,2], data=X_train, x='Annual Income')\naxes[0,2] = 'Annual Income Distribution'\nsns.kdeplot(ax=axes[1,0], data=X_train, x='Health Score')\naxes[1,0] = 'Health Score Distribution'\nsns.kdeplot(ax=axes[1,1], data=X_train, x='Premium Amount')\naxes[1,1] = 'Premium Amount Distribution'","metadata":{"_uuid":"880db4c4-4a9e-4a9c-a7fc-1a9e01b4fbaa","_cell_guid":"b925e22d-12b8-4381-8c69-8fe406292859","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:07.529176Z","iopub.execute_input":"2024-12-18T20:17:07.529475Z","iopub.status.idle":"2024-12-18T20:17:28.784922Z","shell.execute_reply.started":"2024-12-18T20:17:07.529431Z","shell.execute_reply":"2024-12-18T20:17:28.784041Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Vizualization of discrete features","metadata":{"_uuid":"60754b8e-0825-4ded-bf27-13e360be1b58","_cell_guid":"bf136762-6bad-4967-8c5f-f51c2218bfc9","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"numerical_features_discrete","metadata":{"_uuid":"4e380837-46dc-4bb7-8cd2-a68e9cc7721f","_cell_guid":"d556efc7-5ea2-4621-964b-668657526961","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:28.786081Z","iopub.execute_input":"2024-12-18T20:17:28.786462Z","iopub.status.idle":"2024-12-18T20:17:28.792407Z","shell.execute_reply.started":"2024-12-18T20:17:28.786408Z","shell.execute_reply":"2024-12-18T20:17:28.791687Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"fig, axes = plt.subplots(2, 2, figsize=(15,15))\nfig.suptitle('Distributions of discrete features')\n\n\nsns.countplot(ax=axes[0,0], data=X_train, x='Number of Dependents')\naxes[0,0] = 'Number of Dependents Distribution'\nsns.countplot(ax=axes[0,1], data=X_train, x='Previous Claims')\naxes[0,1] = 'Previous Claims Distribution'\nsns.countplot(ax=axes[1,0], data=X_train, x='Vehicle Age')\naxes[1,0] = 'Vehicle Age Distribution'\nsns.countplot(ax=axes[1,1], data=X_train, x='Insurance Duration')\naxes[1,1] = 'Insurance Duration Distribution'","metadata":{"_uuid":"c2567f4b-ec2d-4e67-ba57-6aa5b0323fed","_cell_guid":"d94e649a-c7f9-4c65-bda4-96923a01ff36","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:28.793826Z","iopub.execute_input":"2024-12-18T20:17:28.794515Z","iopub.status.idle":"2024-12-18T20:17:29.952247Z","shell.execute_reply.started":"2024-12-18T20:17:28.794485Z","shell.execute_reply":"2024-12-18T20:17:29.951306Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Vizualization of categorical features","metadata":{"_uuid":"1111df4e-66e3-424b-b90d-27ca5c5a38d8","_cell_guid":"143e8395-23fa-4be1-962a-59562cd05c00","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"categorical_features","metadata":{"_uuid":"73fbfe47-5ba9-4d86-aa00-883df3ffb357","_cell_guid":"218b6ea8-5a2a-40f2-a5bb-166e4e794244","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:29.953554Z","iopub.execute_input":"2024-12-18T20:17:29.954233Z","iopub.status.idle":"2024-12-18T20:17:29.961002Z","shell.execute_reply.started":"2024-12-18T20:17:29.954184Z","shell.execute_reply":"2024-12-18T20:17:29.959978Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"new_array = np.delete(categorical_features, np.where(categorical_features == 'Policy Start Date'))","metadata":{"_uuid":"2170d23a-9e45-408d-b808-d4baf7b8e666","_cell_guid":"86bb2be7-be6d-4177-89e4-bf4aad0511ac","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:29.962148Z","iopub.execute_input":"2024-12-18T20:17:29.962462Z","iopub.status.idle":"2024-12-18T20:17:29.969521Z","shell.execute_reply.started":"2024-12-18T20:17:29.962406Z","shell.execute_reply":"2024-12-18T20:17:29.968676Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"new_array","metadata":{"_uuid":"2d20f541-684f-47d1-9d8e-8d38d8a1df2d","_cell_guid":"cc130e89-e004-42ae-ab94-55c4ae9c56f2","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:29.970599Z","iopub.execute_input":"2024-12-18T20:17:29.970859Z","iopub.status.idle":"2024-12-18T20:17:29.981231Z","shell.execute_reply.started":"2024-12-18T20:17:29.970833Z","shell.execute_reply":"2024-12-18T20:17:29.980447Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def plot_categorical_features(data, features):\n fig, axes = plt.subplots(5, 2, figsize=(15,20))\n fig.suptitle('Distributions of categorical features')\n feat = 0\n for a in axes.flat:\n sns.countplot(ax=a, data=data, x=features[feat])\n a = features[feat]\n feat +=1 \n\n plt.show()\nplot_categorical_features(X_train, new_array)","metadata":{"_uuid":"2a5d63a0-8638-4520-bb2b-eef7976a29ae","_cell_guid":"34d682e7-4668-4930-bf01-366832cc1ef6","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:29.98261Z","iopub.execute_input":"2024-12-18T20:17:29.982853Z","iopub.status.idle":"2024-12-18T20:17:36.50069Z","shell.execute_reply.started":"2024-12-18T20:17:29.982829Z","shell.execute_reply":"2024-12-18T20:17:36.499773Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"Plot value counts for each feature","metadata":{"_uuid":"719c0c8c-2da7-4a21-89de-0c3ac30746e7","_cell_guid":"762fa5f9-d160-4ac1-bd83-749c853a7f2b","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"code","source":"for cat in categorical_features:\n print(X_train[cat].value_counts())\n print()","metadata":{"_uuid":"1badf574-06e1-425a-a952-28ae34fdc06e","_cell_guid":"0c9d124b-9be9-4a88-98bb-132172fd3b29","trusted":true,"collapsed":false,"execution":{"iopub.status.busy":"2024-12-18T20:17:36.501905Z","iopub.execute_input":"2024-12-18T20:17:36.502272Z","iopub.status.idle":"2024-12-18T20:17:37.588087Z","shell.execute_reply.started":"2024-12-18T20:17:36.50223Z","shell.execute_reply":"2024-12-18T20:17:37.587127Z"},"jupyter":{"outputs_hidden":false}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"As we can see categorical features approximately have the almost the same number of categories, except Policy Start Date. I've decided not to plot it, because it would take a lo-o-ot of time.\n\nFrom value counts we can assume types of encoding for our categorical features.\nBinary encoding:\n* Gender\n* Smoking Status\n\nOrdinal encoding:\n* Education Level\n* Policy Type\n* Exercise frequency\n* Customer Feedback\n\nAll other features we can encode by one-hot encoding","metadata":{"_uuid":"2a8e9c94-9e8d-4fa8-8b45-e7e07b98432a","_cell_guid":"a3461cf8-02d4-46d2-a0b3-2b2bd486b818","trusted":true,"collapsed":false,"jupyter":{"outputs_hidden":false}}},{"cell_type":"markdown","source":"## Vizualization of features vs target","metadata":{}},{"cell_type":"markdown","source":"I've did this to see have features depend on target, so I can see which type of relationship is here","metadata":{}},{"cell_type":"code","source":"numerical_features_total","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:17:37.58947Z","iopub.execute_input":"2024-12-18T20:17:37.589888Z","iopub.status.idle":"2024-12-18T20:17:37.596956Z","shell.execute_reply.started":"2024-12-18T20:17:37.589832Z","shell.execute_reply":"2024-12-18T20:17:37.59586Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"for n in numerical_features_total:\n sns.scatterplot(X_train, x=n, y='Premium Amount')\n plt.show()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:17:37.598213Z","iopub.execute_input":"2024-12-18T20:17:37.598537Z","iopub.status.idle":"2024-12-18T20:17:56.839288Z","shell.execute_reply.started":"2024-12-18T20:17:37.598509Z","shell.execute_reply":"2024-12-18T20:17:56.838436Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"## Sort by Premium Amount","metadata":{}},{"cell_type":"markdown","source":"Here I decided to sort dataset by premium amount to get insight which type of users pay the highest premium","metadata":{}},{"cell_type":"code","source":"sorted_data = X_train.sort_values(by='Premium Amount', ascending=False)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:17:56.844305Z","iopub.execute_input":"2024-12-18T20:17:56.844723Z","iopub.status.idle":"2024-12-18T20:17:57.619948Z","shell.execute_reply.started":"2024-12-18T20:17:56.844695Z","shell.execute_reply":"2024-12-18T20:17:57.619205Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"sorted_data.head(10)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:17:57.621182Z","iopub.execute_input":"2024-12-18T20:17:57.62156Z","iopub.status.idle":"2024-12-18T20:17:57.64346Z","shell.execute_reply.started":"2024-12-18T20:17:57.621521Z","shell.execute_reply":"2024-12-18T20:17:57.642478Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import seaborn as sns\n\n# Вибір змінних\nsubset = sorted_data[['Age', 'Annual Income', 'Premium Amount']]\n\n# Візуалізація залежності між віком та здоров'ям для премії\nsns.scatterplot(data=subset, x='Age', y='Annual Income', size='Premium Amount', hue='Premium Amount', palette='cool', alpha=0.7)\nplt.title('Залежність Premium Amount від Age і Health Score')\nplt.xlabel('Age')\nplt.ylabel('Health Score')\nplt.legend()\nplt.grid(True)\nplt.show()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:17:57.644516Z","iopub.execute_input":"2024-12-18T20:17:57.644738Z","iopub.status.idle":"2024-12-18T20:19:00.776441Z","shell.execute_reply.started":"2024-12-18T20:17:57.644715Z","shell.execute_reply":"2024-12-18T20:19:00.775567Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"sns.histplot(np.log(sorted_data['Premium Amount']), bins=30)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:19:00.777735Z","iopub.execute_input":"2024-12-18T20:19:00.778068Z","iopub.status.idle":"2024-12-18T20:19:01.704235Z","shell.execute_reply.started":"2024-12-18T20:19:00.778033Z","shell.execute_reply":"2024-12-18T20:19:01.703357Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"box_features = sorted_data.select_dtypes(include=['object']).columns.tolist()\nbox_features.remove('Policy Start Date')\nbox_features","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:19:01.705564Z","iopub.execute_input":"2024-12-18T20:19:01.706278Z","iopub.status.idle":"2024-12-18T20:19:01.898354Z","shell.execute_reply.started":"2024-12-18T20:19:01.706237Z","shell.execute_reply":"2024-12-18T20:19:01.897477Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"for b in box_features:\n sns.boxplot(data=sorted_data, x=b, y='Premium Amount')\n plt.show()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:19:01.899786Z","iopub.execute_input":"2024-12-18T20:19:01.900094Z","iopub.status.idle":"2024-12-18T20:19:09.139923Z","shell.execute_reply.started":"2024-12-18T20:19:01.900065Z","shell.execute_reply":"2024-12-18T20:19:09.13896Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"X_train.isna().sum()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-18T20:19:09.143744Z","iopub.execute_input":"2024-12-18T20:19:09.144017Z","iopub.status.idle":"2024-12-18T20:19:09.695111Z","shell.execute_reply.started":"2024-12-18T20:19:09.143988Z","shell.execute_reply":"2024-12-18T20:19:09.694208Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}