Решить 10 задач по Питону

Linear Regression Analysis
Use train.csv dataset from Sberbank Russian Housing Market competition (you need to register on Kaggle in order to have an access to its datasets).
Useful links: scikit-learn documentation, in depth: linear regression.
Task 3. Using train.csv dataset, split it in two parts at random (use supporting code below) for train_sample (80% of observations) and test_sample (20% of observations). Perform two linear regression models for price_doc prediction for train_sample (use both numeric and categorical independent variables in your models).
import pandas as pd
import zipfile
zf = zipfile.ZipFile('train.csv.zip')
df = pd.read_csv(zf.open('train.csv'))
from sklearn.model_selection import train_test_split
train_sample, test_sample = train_test_split(df, test_size=0.2)
Task 4. Justify your choice and models' quality with supporting tests, charts and comments briefly.
Task 5. Assess the quality of your models on test_sample dataset. Use RMSE metric. Briefly describe whether or not this quality is appropriate.
K-Nearest Neighbors Analysis
Use any financial ticker (except S&P500) to create the trading strategy, using KNN classifier.
Hint. Use this material to start your own calculations. This link with library fixes might be helpful.
Useful link: scikit-learn documentation.
Task 6. Upload data for one ticker (for the recent year) from Yahoo Finance. Make brief exploratory analysis of obtained data.
Task 7. Apply KNN classifier to obtained data. Assess the quality of the model. Create and briefly justify the trading strategy.
Decision Trees Analysis
Hint. Use this material to start your own calculations.
Useful link: decision trees and random forests.
Task 8. Upload data for one ticker (for the recent year) from Quandl. Make brief exploratory analysis of obtained data.
Task 9. Apply Decision Tree to obtained data. Assess the quality of the model. Create and briefly justify the trading strategy.