0% found this document useful (0 votes)
4 views6 pages

Assignment 2 Data Science

The document is an assignment on Time Series using Python, covering definitions, distinctions between data types, internal components, autocorrelation, and forecasting methods. It includes practical examples such as sales data and calculations for moving averages and exponential smoothing. The assignment also outlines the forecasting process and accuracy measurement techniques.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
4 views6 pages

Assignment 2 Data Science

The document is an assignment on Time Series using Python, covering definitions, distinctions between data types, internal components, autocorrelation, and forecasting methods. It includes practical examples such as sales data and calculations for moving averages and exponential smoothing. The assignment also outlines the forecasting process and accuracy measurement techniques.
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

NAME : SAYANDEEP DUTTA

ROLL NO. : 2302206014


PROGRAM: BBA ACCA ( SEM V)
COURSE : TIME SERIES USING PYTHON
ASSIGNMENT : II
1) Define Time Series. Give two examples of time-dependent
business data.
Ans : A time series is a sequence of observations recorded at regular
time intervals.
Examples:
 Monthly sales of a company
 Daily stock prices

2) Distinguish between cross-sectional and time series data with


examples.
Ans:
Cross-sectional Time Series
Data collected at one point in Data collected over time for one
time across many units unit
Example: income of 100 Example: annual income of one
households in 2024 household for 10 years

3) Explain the internal components of a time series: Trend,


Seasonality, Cyclicity, and Irregularity.
Ans:
 Trend: Long-term upward or downward movement (e.g., sales
increasing over years).
 Seasonality: Regular repeating patterns within a year (e.g.,
festive sales peak every December).
 Cyclicity: Long-term business cycles lasting more than a year
(e.g., recession cycles).
 Irregularity: Random, unpredictable variations (e.g., sudden
strike, pandemic).

4) What is Autocorrelation? How is it measured?


Ans : Autocorrelation measures how a value at time t is related to its
past values (lags).
Measured using the autocorrelation coefficient:
Cov( X t , X t −k )
rk =
Var ( X t )

5) Explain the difference between autocorrelation function (ACF)


and partial autocorrelation function (PACF).
Ans:
 ACF: shows correlation between X t and all lagged values X t −k .
 PACF: shows correlation after removing effects of intermediate
lags — direct effect only.
Usage:
 ACF → identify MA(q)
 PACF → identify AR(p)

6) Explain the difference between autocorrelation function (ACF)


and partial autocorrelation function (PACF).
| Year | 2016|2017 | 2018| 2019|2020|2021 |
| Sales (₹ ‘000)| 40 | 50 | 55 | 60 | 70 | 80 |
Ans : 3-year MA values:
 For 2018: (40+50+55)/3=48.33
 For 2019: (50+55+ 60) /3=55.00
 For 2020: (55+60+ 70) /3=61.67
 For 2021: (60+70+ 80)/3=70.00

7) What is the purpose of data transformation in time series


analysis? Give one example.
Ans :To stabilize variance, remove skewness, or make data more
linear for modeling.
Example: Log transformation Y ' =log ⁡(Y )reduces large fluctuations.

8) Explain Exponential Smoothing. What are its advantages over


moving averages?
Ans : Exponential Smoothing gives more weight to recent data:
F t+ 1=α X t +(1−α ) Ft

Advantages over moving averages:


 Gives more importance to recent observations
 Requires storing less data
 Smooths quickly and adapts to changes

9) A company’s demand in four consecutive months is 100, 110,


120, 115. Using Simple Exponential Smoothing with α = 0.3 and
initial forecast = 100, calculate the forecast for month 5.
Ans : Demand: 100, 110, 120, 115
Initial Forecast: 100
Step-by-step:
 F 2=0.3(100)+0.7 (100)=100
 F 3=0.3(110 )+ 0.7(100)=103
 F 4=0.3(120)+ 0.7(103)=108.1
 F 5=0.3(115 )+ 0.7(108.1)=110.67

10) Describe the Forecasting Process — from data collection to


model evaluation.
Ans:
1. Define objective
2. Collect and clean data
3. Analyze patterns (trend, seasonality)
4. Choose model (moving average, ARIMA, ES, etc.)
5. Estimate model parameters
6. Validate using accuracy measures
7. Make forecasts
8. Monitor and update model

11) Explain how forecast accuracy is measured using MSE or MAD.


Ans:
 MSE (Mean Squared Error):
2
∑ (e t )
MSE=
n

Penalizes large errors heavily.


 MAD (Mean Absolute Deviation):
∑ ∣e t ∣
MAD=
n

Easy to interpret; measures average absolute error.


12) What are the resources and data required for effective business
forecasting?
Ans : Historical data (sales, demand, prices, economic indicators)
 Software/tools (Excel, Python, R, forecasting packages)
 Industry knowledge and expert judgment
 Skilled analysts and computational resources

13) Give one real-life example of time series forecasting (e.g., sales,
rainfall, internet traffic). Explain the type of model used.
Ans : Example: Forecasting monthly electricity demand.
Model used: SARIMA, because electricity demand has:
 Trend
 Strong seasonality (summer/winter peaks)
SARIMA handles both seasonal cycles and non-seasonal patterns.

You might also like