FAM PTT2 SOLUTION

Q1. Attempt any FOUR.

a.Define Reasoning in AI?
Ans: Reasoning in AI is the cognitive process that involves using logic,
knowledge, and information to draw conclusions, make inferences, and solve
problems. It allows AI systems to emulate human-like thinking by analyzing
data, identifying patterns, and making decisions based on the available
information. This process is fundamental to developing intelligent systems
capable of performing complex tasks effectively.
Common Types of Reasoning: - Deductive Reasoning - Inductive Reasoning -
Abductive Reasoning Example of Reasoning: An AI system diagnoses medical
conditions using patient symptoms and a medical database
b.What is Machine Learning?
Ans: Machine learning is a subset of artificial intelligence where computer
systems learn patterns from data to make predictions or decisions without
explicit programming. It uses algorithms to analyze data, identify patterns, and
improve its performance over time, enabling machines to learn from experience
and improve tasks without human intervention.
Types of machine learning:Supervised Learning,Unsupervised Learning.
Examples of machine learning:Social Media, Fraud detection.
c. What is Data Wrangling in ML?
Ans: Data wrangling, also known as data preprocessing, is a crucial step in the
data analysis and machine learning process. It involves cleaning and
transforming raw data into a usable format for analysis. Data wrangling
encompasses several tasks, including data cleaning, variable selection, and data
transformation, all aimed at making the data more suitable for analysis in
subsequent stages. This step is fundamental for obtaining reliable and
meaningful insights from the data.
d.What is Supervised Learning?
Ans: Supervised learning is a machine learning paradigm where the algorithm
is trained on labeled data, meaning the input data is paired with corresponding
output labels or target values. The algorithm learns to map input features to the
correct output by minimizing the error between its predictions and the actual
labels during training.
Algorithms:1.Linear Regression,2.K-nearest neighbours.
Example:Housing Price Prediction.
e.What is First Order Logic and its synatx?
Ans: First-Order Logic (FOL), also known as Predicate Logic or First-Order
Predicate Calculus, is a formal language used for representing knowledge and
making inferences in artificial intelligence and logic. It extends propositional
logic by introducing variables, quantifiers, and predicates, allowing for more
expressive and detailed representation of relationships and properties in a
domain.
Syntax of First-Order Logic: The syntax of FOL consists of various
components: -
Constants: Names for specific objects in the domain (for example, "Alice,"
"Bob"). - Variables: Symbols representing placeholders for objects (for
example, "x," "y").
Quantifiers: - Universal Quantifier ( ∀ ): Indicates that a statement applies to
all elements. - Existential Quantifier ( ∃ ): Indicates that a statement applies to
at least one element.

Q2. Attempt any THREE

a.What are the types of Machine Learning?

Supervised Learning: This type involves training a model on labeled data,

where the desired output is already known. The model learns to predict
outputs from new inputs by finding patterns in the training data.
 Examples include classification and regression.

Unsupervised Learning: In this type, the model is trained on unlabeled

data and must find hidden patterns or intrinsic structures in the input data.
Common techniques include clustering (e.g., k-means) and dimensionality
reduction (e.g., PCA).

Semi-Supervised Learning: This type combines both labeled and

unlabeled data during training. It aims to improve learning accuracy by
leveraging the large amount of unlabeled data along with a small amount of
labeled data. It is often used in applications where obtaining labeled data is
expensive or time-consuming.

Reinforcement Learning: This type involves training a model through trial

and error, where the model learns to make a sequence of decisions by
receiving rewards or penalties. The goal is to maximize cumulative
rewards. Examples include training a robot to navigate a maze or teaching
an AI to play a game.

b.What are the key aspects of Machine Learning Life Cycle?

Ans.1-Gathering Data:-The initial phase of the machine learning life cycle is
Data Gathering, which aims to identify and address all data-related
issues.During this stage, it's crucial to pinpoint the diverse data sources, as
data can be sourced from various outlets such as files, databases, the internet,
or mobile devices. This step holds immense significance in the life cycle, as
the quantity and quality of collected data directly impact the effectiveness of
the final outcomes. Greater data volume translates to more precise prediction.
2. Data Preparation:-Data preparation involves organizing and formatting the
data for its application in machine learning training. During this phase, the
initial step is to consolidate all the data and then randomize its order to avoid
any bias.
Key processes in data preparation:
 Data Exploration:
 Data Pre-processing:
3. Data Wrangling:-Data wrangling, also known as data pre-processing, is a
crucial step in the data analysis and machine learning process. It involves
transforming raw data into a usable format for analysis. Data wrangling
encompasses several tasks, including data cleaning, variable selection, and data
transformation, all aimed at making the data more suitable for analysis in later
stages. This step is fundamental for obtaining reliable and meaningful insights
from the data.
Key aspects of data wrangling:
 Cleaning Data
 Variable Selection:
 Data Transformation
 Handling Missing Values
 Dealing with Duplicates
4. Data Analysis:- The data analysis or machine learning process is the analysis
phase. This phase involves several key activities:
 Selection of Analytical Techniques:
 b.Building Models:
 c.Reviewing the Results:
 d.Iterative Process:
5. Train Model:-Now the next step is to train the model; in this step, we train
our model to improve its performance for better outcomes of the problem.
We use datasets to train the model using various machine learning algorithms.
Training a model is required so that it can understand the various patterns, rules,
and features.
6.Test Model:-After training the machine learning model with a specific
dataset, the next step is model testing.During this phase, the model's accuracy is
assessed by evaluating its performance on a separate test dataset.
7.Deployment:-The final stage in the machine learning life cycle is
deployment, involving the integration of the model into a real-world system.
Prior to deployment, we assess if the model consistently meets performance
criteria and enhances its results with available data.
This phase is akin to delivering the project's final report.
c.List and briefly explain the techniques of knowledge representation
Ans: 1.Logical Representation: Logical representation uses formal logic to
represent knowledge in a precise and unambiguous manner. It follows strict
syntactic and semantic rules to ensure clarity in communication and reasoning.

∀𝑥Human(𝑥)→Mortal(𝑥).
Example: "All humans are mortal" can be represented as
2. Semantic Network Representation: This method represents knowledge as a
network of interconnected nodes (concepts) and edges (relationships), depicting
how various concepts relate to each other.
Example: A semantic network might represent the relationship between
animals and their characteristics, such as "Cat" is a "Mammal," and "Mammals"
have "Fur."
3.Frame Representation: Frames are data structures that hold information
about objects in attribute-value pairs. They provide a structured way to
represent stereotypical situations and facilitate reasoning.
Example: A frame for "Car" might include attributes like "Make," "Model,"
and "Year," with specific values for each attribute.
4.Production Rules: Production rules consist of if-then statements that define
actions based on specific conditions. These rules are used for decision-making
processes in expert systems.
Example: In a medical diagnosis system, a rule might be "If the patient has a
fever and a rash, then consider measles."
d. List and briefly explain the different forms of data.
Ans:- 1. Statistics:- Statistics refers to the mathematical and analytical
technique used to collect ,analyze ,interpret and present data.
It involves methods for summerizing and drawing inference from data ,such as
mean ,median,standard deviation,hypothesis testing and regression analysis.
While it is a fundamental component of data analysis, it is not a form of data
itself but rather a let set of tools for working with data
2.Data Mining:-Data mining is the proces of discovering patterns,
relationships, or useful information from large dataset.It involves techniques
such as clustering, classification,association rule mining, and anomaly
detection.
Data mining is a form of data analysis that focuses on finding valuable insights
within data.
3. Data Analytics:-Daca analytics encompasses the entire process of
examining, cleaning, transforming, and interpreting data to extract meaningful
insights.
It combines statistical analysis, data mining, and visualization to inform
decision-making.
Data analytics is the broader practice of working with data to answer questions
or make informed decisions
4.Statistics Data:-This term appears to refer to data that is relevant so statistical
analysis.
It can include datasets, variables, or information used in statistical studies and
calculations. It's not a distinct category of data but rather data that's used within
the field of statistics.
5. Data Science:-Data analytics is a component of data science.
Data science is a multidisciplinary field that includes data analytics but also
involves other areas such as machine learning, big data technologies, and
domain expertise.
Data science aims to extract valuable insights and knowledge from data to solve
complex problems and make data- driven decisions.
e.Explain Forward chaining and Backward chaining in AI.
Ans:- 1. Forward Chaining in AI:-Forward chaining is an inference or
reasoning strategy used in artificial intelligence to draw conclusions from a
knowledge base by applying rules and facts in a systematic manner. It starts
with the available facts and iteratively applies rules to infer new facts until no
further conclusions can be drawn. Forward chaining is particularly useful for
systems with a large number of potential conclusions and is often used in expert
systems and knowledge-based systems.
A Simple Forward-Chaining Algorithm:
1. Initialization: Start with the known facts.
2. Rule Application: Apply rules whose conditions match the current set of
facts. If the conditions are satisfied, add the rule's consequent (conclusion) to
the set of facts.
3. Repeat: Continue applying rules and updating the set of facts until no new
facts can be added.
Example: Scenario: Diagnosing a medical condition. Process: Start with known
symptoms (e.g., fever, cough). Apply rules to infer possible conditions (e.g., flu,
cold) based on these symptoms.
2 Backward Chaining:-Backward chaining is an inference strategy used in
artificial intelligence to prove the validity of a query or goal byworking
backward from the goal to the available knowledge. It starts with the goal and
uses rules and facts in reverse to find a path to known facts or axioms.
Backward chaining is particularly useful for query-based reasoning and is
commonly used in logic programming languages like Prolog.
A Backward-Chaining Algorithm:
1. Goal Initialization: Start with the goal or query you want to prove.
2. Rule Application: Apply rules whose conclusions match the goal. If the
conclusion matches, try to prove the conditions (premises) of the rule.
3. Recursion: If the premises are goals themselves, recursively try to prove
them using the same backward-chaining process.
4. Base Case: Eventually, the process will reach known facts or axioms that do
not need further proof.
Example:
Scenario: Verifying a hypothesis.
Process: Start with the hypothesis (e.g., the patient has the flu). Work backward
to find supporting symptoms (e.g., fever, cough) and check if these known facts
are present.