Question 1:
Explain different approaches used for analysis.
Answer:
The data analytics maturity model categorizes the various manners of analyzing the
data by their primary goal (descriptive, diagnostic, predictive, and prescriptive) and
by research methodology (qualitative, quantitative, and mixed-methods) at a very high
level. Nevertheless, the particular methods are subject to the type of data and the
required insights.
Analytics Types:
The extent and intensity of these four basic methods complement one another thereby
generating insights and aiding decision-making.
Descriptive analytic methods:
Provide historical data summaries aiming to clarify the course of events leading up to
the present, or the present situation while at the same time pointing out the methods of
central tendency description (mean/average, median, mode) and loosely displaying
the methods of the dispersion examination description (variance, standard deviation).
Descriptive analysis methods gain support from the common graph and chart
visualizations. Descriptive analytical methods give a high-level view of a
characteristic or performance.
Diagnostic Evaluation:
This process lets you dive deeper into the data and discover "what influenced the
occurrence of an event". It applies descriptive analysis results to investigate the
correlation between them and possible causative factors - searching for patterns or
exceptions. Among the techniques that can help to get the root cause are regression
analysis, mining data, and applying probability theory.
Predictive analysis:
With the help of statistical models, machine learning techniques, and past data,
predictive analytics forecasts future happenings. It supports trend spotting which
could result in sound predictions; nevertheless these predictions are theoretical.
Regression models, time-series methods, and Monte Carlo simulations are only a few
of the many well-known examples.
Prescriptive Analysis:
The last phase of the process where all the previous analyses' inputs and outputs are
merged together to provide an answer to the question of which step to take for the best
results. It uses certain optimization and simulation algorithms to deliver practical and
individual steps and routes to take; for instance, recommending the best way to
deliver a parcel by considering traffic and climate conditions.
Methods and Techniques:
Qualitative vs. Quantitative:
In general, the primary methodology in the analysis was determined by the data type.
Quantitative Analysis:
This technique makes use of numeric data and is mainly concerned with discovering
statistical relationships, patterns and trends.
Most Used Methods:
Statistical Analysis:
Inferential statistics methods include assessment of median t-tests, ANOVA, and Chi-
square tests among others for sample based on which inference is made to the
population.
Regression Analysis:
It looks at relationships between variables to determine influence and make
predictions.
Cluster Analysis:
This divides the like data points into unique clustering and also detection of segments
or organization within the data.
Time Series Analysis:
This uncovers trends, seasonality and cycles in the data collected over a period of
time.
Question 2 :
Differentiate between the differential and integral formulation for problem
solving.
Answer:
Differential Formulation Integral Formulation
Expresses the problem in terms of rates of Expresses the problem over a region or
change (derivatives). domain using integrals.
Describes local behavior of a variable at a Describes global/overall behavior over an
specific point. entire area or volume.
Governs differential equations such as Uses conservation laws (mass, momentum,
Differential Formulation Integral Formulation
Laplace, Navier–Stokes, etc. energy) in integral form.
More general and often easier for complex
More detailed and exact representation.
geometries.
Example:
describes bending of a Example: gives total force
beam at a point. on a surface.
Engineering Example
Differential Form: Heat conduction in a wall described by
2
d T
2
=0.
dx
Integral Form: Total heat transfer using
dT
Q= ∫ kA dx .
dx
Question 3 :
Describe Analytical, Numerical and Experimental approaches used in
Engineering.
Answer:
In general, engineering resolves issues through one of three methods:
analytical, numerical, or experimental. It will rely on each of them to confirm (verify)
and also supply the others with information.
Analytical Approach Definition:
The analytical approach is the one that applies fundamental mathematical theories and
symbolic manipulation to derive exact, closed-form answers using a "pencil and
paper" technique. It is the case in which it is possible to solve directly for unknown
variables and is mostly based on physical laws (e.g. Newton's laws, etc.) as well as
mathematical equations.
Characteristics:
The method yields exact (100 percent) results for the provided input values. One of
the major benefits will be the possibility of acquiring a thorough knowledge of the
factors that affect the results in varying degrees. Usually, it is a walk through the
relatively simple applications only which are in an uncomplicated, idealized and
linear form; whether it be a problem or a system with a standard geometry.
Some typical applications are:
The calculation of the deflection of a simply supported beam using the basic beam
bending equations, or the prediction of fluid flow in a simple pipe by using a
mathematical formula.
Numerical Approach Explanation:
The numerical approach consists in transforming a complex mathematical problem
into a series of arithmetic and logical operations that can be handled by computational
methods. Approximate solutions are obtained by numerical methods through the
discretization of the problem area into small elements (e.g., in Finite Element
Methods) and iteratively solving a large system of equations with the computer's
assistance.
Examples are the use of CFD to predict the airflow over an automobile or the use of
finite element analysis for the stress in a complicated structural component.
Experimental Methodology:
Definition:
So-called physical testing, or experimental methodology is a method usually done on
solid or physical models, both under controlled and unrestrained situations. This
method makes use of physical observation and measurement to monitor physical
occurrences.
Characteristics:
Gives very comprehensive and trustworthy data since the real system's performance is
actually being measured. Essential for confirming the hypotheses of the analysis and
studies of the numerical model. Experimental methodology can turn out to be
incredibly costly and time-consuming requiring the physical prototypes, necessary
equipment, and testing facilities. In addition, we should be aware that the results
might entail measurement errors, imperfections in the equipment, and all practitioners'
limitations.
To mention a few:
The wind tunnel testing of an aircraft model to get the lift and drag coefficients, or the
installation of strain gauges on a bridge to quantify the actual stress in the bridge due
to traffic load.
Summary of Differences
Feature Analytical Approach Numerical Approach Experimental
Approach
Solution Exact, closed-form Approximate (via Real-world
Type algorithms) measurements
Complexity Limited to simple problems Can handle highly Applicable to real
complex problems systems
Tools Pen, paper, calculator Computers, specialized Physical prototypes,
software measuring equipment
Cost/Time Generally quick and Varies, can be Expensive and time-
inexpensive computationally consuming
intensive
Primary Fundamental understanding, Complex design, Validation, physical
Use simple design, validation simulation, proof, data collection
basis optimization
Question 4:
Explain significance of Taylor’s Series.
Answer:
Taylor's series is significant because it provides a way to approximate complex
functions with simpler polynomials, using only the function's value and its derivatives
at a single point. This approximation is a powerful tool in calculus and applied
mathematics for solving problems where a function's exact form is difficult or
impossible to work with, and its accuracy improves with more terms. This method is
crucial for tasks like numerical calculations, solving differential equations, and
understanding function behavior near a specific point.
Key significances of Taylor's series
Function approximation: It allows any sufficiently smooth function to be
represented by an infinite sum of polynomials, which can be truncated to
create approximations (Taylor polynomials). The more terms you include, the
more accurate the approximation is near the center point.
Local analysis: The series provides a method to understand a function's
behavior (like its slope, curvature, etc.) in a local region around a specific
point, based only on the function's derivatives at that single point.
Simplification: It transforms complex functions into a series of polynomial
terms, which are easier to differentiate, integrate, and manipulate algebraically.
Numerical computation: Truncated Taylor series are used to perform
numerical calculations for functions like sine, cosine, and exponentials, which
is fundamental in many algorithms for scientific computing and engineering
applications.
Problem-solving: It is used to find solutions to differential equations and to
analyze systems that are not linear by linearizing them around a point.
Foundation for other math: It connects to and is used in other areas of
mathematics, such as complex analysis, to extend the properties of analytic
functions.
Question 5.
Explain deterministic and probabilistic approaches used in mathematics.
Answer:
deterministic approaches assume that the outcome of a system or model is entirely
predictable from its initial conditions and parameters, with no room for randomness.
In contrast, probabilistic approaches incorporate elements of chance and uncertainty,
expressing outcomes as a range of possibilities with associated likelihoods.
Deterministic Approaches
Deterministic models function based on fixed rules and precise, known inputs,
ensuring that the same input always produces the exact same output.
Characteristics:
o Predictable: The future state of the system is fully determined by its
current state.
o Transparent: The logic and rules are clear and easily auditable,
making the decision-making process traceable.
o Requires complete data: They rely on accurate and complete input
data; missing or inconsistent data can cause the model to fail.
o Output: Produces a single, precise outcome or "point estimate".
Examples and Applications:
o Physics and Engineering: Predicting the motion of a pendulum using
classical mechanics equations, assuming no air resistance or friction
variability.
o Finance and Operations: Simple calculations like payroll costs or
inventory levels based on known demand and lead times.
o Computer Science: Standard algorithms like sorting or a deterministic
Turing machine where each step is explicitly defined and yields
consistent results.
o Queuing Models: Assessing a system where customer arrivals and
service times are known with certainty.
Probabilistic Approaches
Probabilistic (also known as stochastic) models are designed to handle inherent
uncertainty and variability in real-world phenomena by using probability theory and
statistical inference.
Characteristics:
o Uncertain Outcomes: Outcomes are expressed as a probability
distribution over possible results, not a single value.
o Adaptable: Can learn and adapt from new data and are better suited
for evolving or dynamic environments.
o Tolerates incomplete data: Can function effectively with incomplete
or inconsistent data by calculating likelihoods.
o Output: Provides confidence scores or a range of potential outcomes.
Examples and Applications:
o Weather Forecasting: Models that predict the probability of rain or a
range of potential temperature outcomes.
o Machine Learning: Spam filters that assign a probability (e.g., "85%
chance this is spam").
o Risk Assessment: Simulating a range of potential disaster scenarios
and their likelihoods to quantify potential losses and inform risk
management strategies.