Download as doc, pdf, or txt
Download as doc, pdf, or txt
You are on page 1of 9

CHAPTER THREE

SYSTEM ANALYSIS AND METHODOLOGY

3.1 System Analysis

Rainfall prediction is essential for agriculture and water resource management. The method is
based on Artificial Neural Network (ANN) as a proposed model to forecast the amount of
rainfall in a particular region. Historical weather data and climatic variables such as temperature,
humidity, wind speed, and air pressure are used as inputs. This method was described by Smith
et al., (2019).

3.2 Review of Existing System

The current system is concerned with predicting rainfall by using traditional statistical methods
to analyze historical weather data. However, this approach may not be accurate enough to meet
the requirements of agriculture and water resource management. In another study, researchers
used Artificial Neural Network (ANN) techniques on a set of rainfall data obtained from the
field. This approach can help farmers and water resource managers to plan and manage their
operations more effectively. By comparing different ANN models, the researchers can determine
which model provides the most accurate rainfall prediction. This can help reduce stress on
farmers and water resource managers and enable them to make informed decisions. This method
was described by Gupta et al., (2020).

3.2.2 Problems with Existing System

Some of the problems faced with the existing system for rainfall prediction are outlined below:

i. Due to the complex nature of climate and weather patterns, it is challenging to


anticipate rainfall accurately in real-time.

ii. The focus has been on using traditional statistical methods to analyze historical
weather data rather than exploring alternative data preprocessing and feature selection
techniques that can prepare the dataset more suitable for Artificial Neural Network
(ANN) models.
These problems can lead to inaccurate rainfall predictions, which can have significant
consequences for agriculture and water resource management. Therefore, it is essential to
explore new approaches such as ANN and investigate the optimal data preprocessing and feature
selection techniques that can improve the accuracy of rainfall predictions. This can help farmers
and water resource managers make better-informed decisions and reduce the risks associated
with weather-related events.

3.3 Review of the Proposed System

The proposed method combines computer science and statistics to provide a strategy that will
make rainfall prediction more accurate and reliable. The system proposes to use an Artificial
Neural Network (ANN) model, which is a machine learning technique, to predict rainfall. The
ANN model will use historical weather data and climatic variables as inputs to generate new data
that can be used to predict the amount of rainfall in a particular region.

By using machine learning techniques, the proposed method can learn from past data and
improve its accuracy over time. This can help farmers and water resource managers to make
better-informed decisions and mitigate the risks associated with weather-related events.
Additionally, the proposed method can reduce the burden on human experts who are responsible
for analyzing and interpreting weather data. This method was described by Kumar et al., (2018).

3.4 Justification for the Proposed System

The justification (advantages) of the proposed system for rainfall prediction is as follows:

1. The system has the potential to significantly improve the accuracy of rainfall prediction,
which can help farmers and water resource managers to plan and manage their operations
more effectively.

2. By reducing the uncertainty associated with rainfall prediction, the proposed system can
help mitigate the risks associated with weather-related events, which can have significant
economic and social consequences.
3. The system provides users with an easy-to-use interface to access rainfall predictions,
which can help increase awareness and educate both old and young on the importance of
taking rainfall prediction seriously.

4. The proposed system can serve as a valuable tool for government agencies, disaster
management organizations, and other stakeholders involved in managing natural
resources. By providing accurate rainfall prediction, the proposed system can help these
organizations to make informed decisions that can benefit society as a whole.

Overall, the proposed system has the potential to address some of the critical challenges
associated with rainfall prediction and can help improve the resilience of agricultural and water
resource systems.

3.5 Functional Requirement

I. The functional requirement describes or explains the system's behavior; it defines the
software system's functionalities. They also specify what the proposed system is intended
to do or, more precisely, the services that the system is supposed to provide to its users.

II. The following are the functional requirements for the proposed system:
prediction/forecasting of rainfall using an Artificial Neural Network (ANN) will necessitate
the following steps:

III. Collect the historical weather data and climatic variables that are related to rainfall
prediction in a particular region.

IV. Pre-process the dataset to remove any missing values, outliers, or irrelevant features that
could affect the accuracy of the ANN model.

V. Use the pre-processed data to train and test the ANN model.

VI. The trained ANN model can then be used to generate new data that can be used for rainfall
prediction.
In this research study used, Object-Oriented Analysis and Design (OOADM) methodology is
followed to achieve the task. The proposed system aims to provide accurate and reliable rainfall
prediction to help farmers and water resource managers make better-informed decisions. The
functional requirements outlined above provide a roadmap for developing the system and
achieving this objective

3.6 High level Model of the Proposed System


The general description/architecture of the program for the proposed rainfall prediction system
using an Artificial Neural Network (ANN) is known as High-Level Design (HLD). It comprises
a description of the system architecture, database design, and a summary of the systems,
services, platforms, and module relationships.
The following diagram below depicts the HLD of the proposed system:

Rainfall Prediction
Model

Dataset Performance

Learning

No. of Days Model Prediction

Training

Testing
Figure 3.1 High Level Model for Rainfall Prediction System

The system architecture consists of three main modules: data collection module, pre-processing
module, and ANN model module.
The data collection module is responsible for collecting historical weather data and climatic
variables related to rainfall prediction in a particular region. This module communicates with a
database that stores the collected data.
The pre-processing module is responsible for cleaning the collected data, handling missing
values, outliers, and irrelevant features that could affect the accuracy of the ANN model. This
module also communicates with the database to store the pre-processed data.
The ANN model module is responsible for training and testing the ANN model with the pre-
processed data. This module uses an ANN algorithm to generate the predicted rainfall data that
can be used for decision-making.
The system architecture also includes a user interface module that allows users to input relevant
data, select the region of interest, and view the predicted rainfall data. The user interface module
communicates with the ANN model module to retrieve the predicted rainfall data and display it
to the user.
Overall, the proposed system architecture provides a framework for developing an accurate and
reliable rainfall prediction system that can benefit farmers and water resource managers.

3.7 Data Flow Diagram of the Proposed System


A Data Flow Diagram (DFD) is a graphical representation of the flow of data within a system or
process. Unlike other diagrams, DFDs do not include control flow, loops, or decision rules. They
primarily illustrate the inputs, outputs, and processes involved in each entity, along with relevant
information about them. In this study, the proposed system design and rainfall prediction will be
analyzed using the DFD approach. The data flow diagram of the proposed system is presented in
figure (3.1).
Load Rainfall
dataset

1 Provide the 3
data
Discovery Rainfall data
Pattern prediction module
Extract knowledge based
patterns

Analyze
expected
input
Train

2 4
actual: value,
predicted: value
Model Training Predicted analysis
Module Module

Display Predicted Outcome of result


Result Database

Figure 3.2 data flow diagram Model for Prediction System


3.8 Research Methodology
Object-Oriented Analysis and Design (OOADM) aims to establish a plan for software
development. It involves using UML and the Unified Process to define object-oriented analysis,
where data manipulation processes are encapsulated in objects and subsystems are divided into
modules. OOADM is preferred for the proposed system because it is best suited for modeling
both static and dynamic semantics. Desktop applications are also ideal for OOADM.

3.8.1 Proposed Algorithm

Artificial Neural Network (ANN) is a machine learning method that uses a network of
interconnected nodes to process information. ANN is well-suited for prediction tasks, such as
rainfall prediction, by analyzing historical data patterns and using them to forecast future values.
The proposed system design for rainfall prediction using artificial neural network involves the
following steps:

Step One: Data Collection

The first step is to collect and prepare the rainfall data set. This includes gathering historical
rainfall data, selecting relevant attributes, and preprocessing the data to remove noise and
inconsistencies.

Step Two: Feature Extraction

In this step, relevant features are extracted from the preprocessed data to create a feature vector
that represents the rainfall patterns.

Step Three: Training and Testing

The feature vector is then used to train and test the artificial neural network. The network is
trained using a set of labeled data, and tested using a separate set of data to evaluate its accuracy.

Step Four: Results

The final step involves analyzing the results of the trained neural network to make rainfall
predictions for future time periods.
This approach, which utilizes artificial neural network for rainfall prediction, is a promising
solution due to its ability to handle complex patterns and large amounts of data.
3.8.2 Artificial Neural Network (ANN) Approaches
Artificial Neural Network (ANN) is a supervised learning method used for classification and
regression. ANN comprises multiple interconnected processing nodes that work collectively to
provide a prediction or decision based on the input data. The network is trained using a set of
input-output data pairs, which enables it to learn patterns and relationships in the data. Unlike
Support Vector Machines, ANN can handle non-binary data, and the output can be continuous or
discrete. The network can also be designed to have multiple layers, allowing it to learn and
extract complex features from the input data.
3.8.3 Data Set Description
In the present study, a forecasting technique model is developed to predict the amount of rainfall
using an artificial neural network. The model analyzes existing historical sample data sets and
generates a statistical report that can be used to make future predictions. Various factors that
affect the amount of rainfall are factored into the model, and the training and testing data sets are
extracted from an online repository. The artificial neural network is utilized to develop the
model, which can be used in a decision-making support system. The dataset contains hourly
rainfall data with their respective attributes.

3.8.4 Feature Set Description


Initially feature set description will base on condition of attribute which need to be centered on
the main subjects and have to be taught theoretically. The attributes along with their descriptions
and possible values will be presented in an excel format in order to be used for further analysis.
3.8.5 Experimental Setup
First, we can import the necessary libraries such as pandas for data manipulation, scikit-learn for
machine learning tasks, and matplotlib for visualization.

Next, we can load the dataset into a pandas dataframe and perform data preprocessing steps such
as removing missing values and scaling the features.

After that, we can split the dataset into training and testing sets using scikit-learn's
train_test_split function.
Then, we can perform feature selection using scikit-learn's feature selection algorithms.
Finally, we can train and test different machine learning models such as artificial neural
networks for rainfall prediction and evaluate their performance using metrics such as accuracy,
precision, and recall.
All of these steps can be implemented and documented in a Jupyter Notebook using Python code
cells and markdown cells for explanations and visualization.
Step 1: Load the historical rainfall dataset into the model for future analysis.
Step 2: Preprocess the data and split it into training and testing sets.
Step 3: Select the attributes for prediction analysis from the given dataset.
Step 4: Initialize the artificial neural network with appropriate parameters such as the number of
layers, nodes, activation functions, and learning rate.
Step 5: Train the neural network using the training dataset and backpropagation algorithm.
Step 6: Validate the accuracy of the trained model using the testing dataset.
Step 7: Use the trained neural network to predict future rainfall values based on new input data.
Step 8: Store the predicted values in an output array.
Step 9: Repeat steps 7-8 until all the data has been predicted.
Step 10: Analyze the results and calculate the accuracy of the model using metrics such as mean
absolute error or mean squared error.
Step 11: Adjust the neural network parameters as necessary to improve the accuracy of the
model.

You might also like