Table of contents

• Introduction
• Calling data from web
• Data
• Setup

Introduction

This work is about estimating the price of a stock in an interval that has been affected by price manipulation. Therefore, we want to apply some statistical approaches to estimate the price of a stock using information from other stocks.

Calling data from web

By using the following scripts, we can download information about stock prices in Iran

import numpy as np
import pandas as pd
import pytse_client as tse
import jdatetime , glob, os, ntpath
from datetime import datetime
from pytse_client import download_client_types_records

tickers = tse.download(symbols="all", write_to_csv=True,  include_jdate=True)

After executing the above code, the daily transaction information of all stocks will be downloaded and stored in the code directory. Now, we combine all data together:

path = r'C:\\Users\\masoumifard.kh\\CodsOfprojects\\Public\\PredictionOfPumpAndDump\\tickers_data'
filenames = glob.glob(path + "/*.csv")
def path_leaf(path):
    head, tail = ntpath.split(path)
    return tail or ntpath.basename(head)
Name = [path_leaf(path) for path in filenames]
dfs = []
for filename in filenames:
    dd = pd.read_csv(filename)
    strs = [path_leaf(filename).split(".")[0] for x in range(len(dd))]
    dd.insert(1,"PersianNameOfCode", strs, True)
    dfs.append(dd)
big_frame = pd.concat(dfs, ignore_index=True)

Because of the dependence between daily stock prices, we have to work with the rate of return instead of the stock price.

ReturnRate = (big_frame['adjClose']-big_frame['yesterday'])/big_frame['yesterday']
big_frame.insert(1,"ReturnRate", ReturnRate, True)