INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 993
www.rsisinternational.org
Planting Sustainability Prediction Using Random Forest Algorithm
Harold R. Lucero
1
, Luisito DC. Soriano
2
, Jayson D. Dela Cruz
3
, Rodel I. Magniyo
4
, John Marvic G.
Hernaez
5
1,2,3
College of Graduate Studies, Quezon City University
4
Responsive System Division, Office of the President
5
Corporate Planning-Data and Analytics, Century Pacific Food Inc
DOI:
https://doi.org/10.51244/IJRSI.2025.1210000089
Received: 20 October 2025; Accepted: 26 October 2025; Published: 05 November 2025
ABSTRACT
This study aimed to utilize smart space technologies in determining planting sustainability and soil moisture
stress that will optimize IoT and data mining technique to help in achieving the government’s goal of
empowering and strengthening the nation’s agricultural sector. The study employed the combined
experimental, developmental, and quantitative research approach towards achieving the objectives of the
study. In order to gather abiotic and edaphic data, including soil moisture, light, temperature, humidity, pH,
and NPK level, the study constructed an IoT prototype. The researcher employed LoRa technology in
transmitting collected data into the IoT gateway before uploading it to Firebase Realtime Database. The study
also involved the development of a mobile application using Blynk IoT Framework and web interface for
remote monitoring and control of irrigation and UV lighting system. Comparative analysis was conducted
between Random Forest, KNN, and Naïve Bayes Algorithm in predicting planting sustainability based on
available data. Based on the calculated Kappa of 0.9901, Random Forest demonstrated the highest level of
accuracy with a "Almost Perfect" strength of agreement. Random forest was implemented using the RubixML
library to enable the web interface to perform predictions of Planting Sustainability on data stored in the
database. In the user evaluation test based on ISO25010 conducted by the researcher, the overall weighted
mean for all criteria is 4.19, with an "Agree (A)" interpretation. This indicates that the developed system is of
excellent quality, excelling in functionality, dependability, usability, efficiency, maintainability, and
portability.
Index Terms: IoT, Data Mining, Agriculture, Vertical Farming, Wireless Sensor Node, Arduino, Blynk
INTRODUCTION
The Philippines is recognized as an agricultural nation since agriculture contributes significantly to the
country's economy and because 32% of its total geographical area is made up of rural areas. [1]. The
agricultural sector's Gross Domestic Product (GDP) serves as a measure of the agricultural sector's growth,
which is essential to the nation's economic situation[2][3]. According to the data from the World Bank,
Philippines's average GDP is 21.36 percent from 1960 to 2016. Based on records, the highest percentage was
31.06 percent in 1974, while the lowest was 9.65 percent in 2016. This demonstrates the unfortunate reality
that agriculture's contribution to GDP (value added) is declining[4][5].
In order to increase and improve the productivity of the agriculture sector, the Philippine Department of
Agriculture (DA) is increasingly embracing numerous technology innovations. Additionally, DA is exploring
the use of "smart space" technologies to enhance farmers' access to crucial information about plants, soil
conditions, and climate state, thereby enhancing their competitiveness in high-value agro-products[6][7]. This
is also done to address the challenges of the future, such as tripling the amount of food produced and
maintaining food security, commercial margins, and climate change mitigation[8][9]. To meet this demand,
stakeholders are currently embracing the concept of internet of things (IoT) for analytics and increased
production capacity [10][6]. IoT refers to interconnected objects, tools, vehicles, structures, sensors, and
computer applications that are used to establish communication and exchange data. IoT enables users to track
and regulate the environment remotely using an existing network architecture. The physical environment can
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 994
www.rsisinternational.org
be linked with computer technology and online virtual resources to provide end users with both valuable
information and functionalities[11][12]. The IoT is going to advance farming's future. Smart farming is already
becoming more popular in the agricultural industry as a result of technological advances, and smart farming is
rapidly becoming a norm[2].
The basic concept is to boost lettuce growth while maintaining high quality in a controlled environment [7] or
residential area by employing vertical farming to maximize the use of available land. Vertical farming is an
agricultural technique where crops are cultivated by stacking them vertically.
Successful agricultural production depends on the availability of the proper amounts of soil moisture, which is
also important for evaluating the sustainability of planting. Lack of moisture may harm plants and reduce
production. Root disease and wasted water are the results of excess[13]. Water serves as a delivery system for
any nutrients that are not strongly linked to the soil, which is equally crucial. Therefore, it is essential to ensure
that crops receive the optimum amount of water supply at the correct point of time, where understanding the
level of soil moisture is crucial [14][15].
In view of this study, a data mining technique using a classification algorithm will be employed to determine
planting sustainability. This is to make predictions on planting sustainability based on climatic variables so that
proper intervention can take place to ensure the high quantity and quality of the harvest yield.
This study suggested a solution that will make use of the IoT concept by developing an Arduino-based system
that can track essential parameters in real time and use a classification algorithm to predict the planting
sustainability. The system will also activates the irrigation system if the level of soil moisture falls below a
predetermined threshold to avoid water scarcity.
Related Studies
According to the author, the goal of experimental research is to draw conclusions from the application of
theories in the real world and to test the validity of those conclusions. This CS technique is employed in a
variety of disciplines, including artificial neural networks, theorem-proving automation, natural language
processing, performance and behavior analysis, etc. It is crucial to reiterate that all experiments and findings
must be able to be replicated. If we apply the experimental approach in the field of IS, we could also need to
use other techniques or instruments. These techniques or instruments are utilized to back up and demonstrate
the validity of the developed project. The technical issue is the first area of the project that can be evaluated
using success criteria for the established project, such as availability, dependability, scalability, and stability.
The second area of testing that requires input from system users is the project's usability; findings for this area
can be achieved by the statistical analysis of a questionnaire, a tool that is utilized in conjunction with the
experimental approach[16].
A published study by S. Dutta et al., proposed a way to technologically manage the entire procedure of
assessing requirements to maintain plant development and sustainability such as soil moisture, temperature,
relative humidity, and light without requiring the presence of growers[17].
In their article [24], Vimal and Shivaprakasha defined greenhouses as a regulated environment where plants
can grow. The constant monitoring and management of ecological parameters are crucial for a greenhouse
system to achieve the most extreme plant development. The researchers also provided formulae, which are
illustrated below, for computing various climate variables based on voltage values obtained from DHT11 using
an Arduino microcontroller.
Harshani et al. identified soil properties and factors such as texture, structure, porosity, and color in their
research. The ratio of soil particles to organic matter is referred to as texture, structure is the aggregation of
primary soil particles, porosity is the number of pores in the soil, and color is determined by its organic content
and oxidation degree. Temperature, humidity, soil moisture, and pH level are all important soil parameters for
farming. Temperature is an important component in plant growth, while humidity refers to the amount of water
vapor in the soil. Soil moisture is required for plant nutrient uptake, and pH levels should be between 5.5 and
7.5 [18].
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 995
www.rsisinternational.org
Researchers from India presented their findings in 2023 along with a solution they think will deal with the
scarcity of arable land and water availability. The researchers introduce the idea of vertical farming. In order to
grow plants vertically, layers of plants are typically stacked indoors. Vertical farming does not depend on the
weather, utilizes less land, and produces more yield per square meter[19].
Embedded systems are used to enhance a set of smart applications running in a physical environment known as
a "smart space." The study highlighted adaptability, interoperability, openness, expandability, and self-
management as qualities of a smart space. [20].
A study was carried out in Porto, Portugal, with the goal of employing smart cities to enhance sustainability
and quality of life. The UrbanSense project was used by the researchers to gather information on weather
patterns, quality of life, and air quality. Additionally, researchers created an ontology for describing data from
smart cities, which UrbanSense researchers then validated. All data were then published in internal data
repositories for big data and research purposes. [21].
The authors used the Smart Space application to tackle the problems of meaningful event recognition, context
information gathering, and visualization in an event recording system. This enables the system to visualize
events using a timeline based on readings from standalone sensors or sensor networks and actions done by the
smart space application[22].
The authors used the smart space idea to assist in managing equipment, decrease repetitive behavior, and
provide convenience by acknowledging and implementing the recognized user's configuration automatically.
The system was intended to recognize and monitor users in the smart room by remembering their usual
activities such as switching on the light and TV and automatically saving it in the configuration of the user[23].
Similar research was conducted by the authors to utilize the concept of smart space technology to turn a
conventional home into smart homesto aid and give comfort to physically disabled people. The proposed
project is eyeing the idea of integrating digital assistants with the available smart devices in the markets today.
This makes it possible for people with disabilities to manage their homes by controlling appliances with voice
commands through a variety of readily available voice assistants [24].
The Internet of Things (IoT) is a network of connected objects connected to a cloud server. Embedded
technology significantly impacts IoT by enabling the development of smaller, energy-efficient, and high-
performance controllers that offer effective data storage, receiving, and control in their actuators [25].
DESIGN AND METHODOLOGY
Research Design
A research design includes the structure of a study and the strategies for conducting the studies. The study
utilizes the combined experimental, developmental and quantitative research approach towards achieving the
objectives of the study. This research approach is suitable for this study since it involves the development of
the proposed prototype; develop a model; and evaluate the system’s compliance in software development
standards.
Fig. 1 Research Design of the Study
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 996
www.rsisinternational.org
Figure 1 shows the research design of the study. The study's research design includes prototyping for
developmental research, data mining for experimental research to predict planting sustainability, and
quantitative research for determining compliance with ISO25010 standard.
Developmental Research Design
Fig. 2 Prototyping Model [26]
This study employs a prototyping model in the developmental research design. The Figure below illustrates the
process and the expected outputs in every phase of the prototyping model.
1 ) Requirements:
In this stage, the researcher carefully analyzed the requirements and specifications needed in the development
of the prototype that will help satisfy the research objectives. Different system functionalities as well as
hardware components were identified. Abiotic and edaphic elements including temperature, soil moisture, light
intensity, humidity, pH, and NPK level are among the abiotic and edaphic variables that the prototype is
designed to collect data on. Data collected will be uploaded to a cloud server and later be used for identifying
an algorithm that will determine planting sustainability.
2 ) Quick Design:
Fig. 3 Block Diagram of the Prototype
During this phase, an initial design was created. Fig. 3 illustrates how each component will interact and how
data will flow from one component to another. The Wireless Sensor Node (WSN), IOT Gateway, Irrigation
and UV Lighting System, Web Interface, and Mobile Application that displays the reading of various sensors
are some of the components that make up the prototype. The mobile application also allows end-users to gain
control of the irrigation and UV lighting system remotely.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 997
www.rsisinternational.org
3 ) Build Prototype:
Fig. 4 System Architecture Design
Fig. 4 illustrates the architecture design that guides the researcher in building the prototype.
4 ) User Evaluation:
In this stage, the IT experts and agriculturists were invited to evaluate and to assist with discovering the
strengths and weaknesses of the prototype. Comments and ideas were gathered from all the participants to
further enhance and fix the errors discovered. The researcher adopted ISO25010 to evaluate the prototype. The
prototype was evaluated based on characteristics stipulated in ISO25010 such as functionality, performance
efficiency, compatibility, usability, reliability, security, maintainability, and portability [27]. Weighted mean
responses from respondents were used to determine the prototype's conformity with the ISO standard. The
researcher applied weighted mean as a statistical tool for interpreting the data gathered from the responses
given by the respondents to evaluate the prototype. The formulas presented assessed the performance of the
system.
Where:
W - Weighted Mean
n - number of terms to be averaged
wi - weight applied to x values
Xi - data values to be averaged
Equation 1
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 998
www.rsisinternational.org
Table I Five-Point Likert Scale
Numerical Value
Range
Interpretation
5
4.51 5.00
Strongly Agree (SA)
4
3.51 4.50
Agree (A)
3
2.51 3.50
Moderately Agree (MA)
2
1.51 2.50
Disagree (D)
1
1.00 1.50
Strongly Disagree (SD)
Table I shows the Likert scale used in evaluating the prototype. The numerical value of 5 with a scale of 4.51
to 5.00 is interpreted as Strongly Agree (SA). The numerical value of 4 with a scale of 3.51 to 4.50 is
interpreted as Agree (A). The numerical value of 3 with a scale of 2.51 to 3.50 is interpreted as Moderately
Agree (MA). The numerical value of 2 with a scale of 1.51 to 2.50 is interpreted as Disagree (D). Lastly, the
numerical value of 1 with a scale of 1.00 to 1.50 is interpreted as Strongly Disagree (SD).
5 ) Refining Prototype:
Tests such as unit testing and integration testing were regularly conducted during the construction of the
prototype. All bugs found during tests need to be corrected to enable the prototype to read and generate
accurate results. The researcher conducted a user evaluation to improve the prototype and identify errors
impacting accuracy and efficiency. IT experts are being consulted to assess and provide suggestions for the
project's improvement. This evaluation will help identify areas for improvement and enhance the project's
overall performance. During this stage, all ideas and comments from the stakeholders were collected and
addressed. Different alternatives were tested to meet the requirements and implement the changes.
6 ) Implement and Maintain:
The researcher built a wall-mounted platform to deploy the created prototype because there are no areas for
performing conventional horizontal farming. This platform embraces the idea of vertical farming and may be
utilized by residents interested in urban farming.
Fig. 5 Wall-Mounted Platform for Vertical Farming
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 999
www.rsisinternational.org
To prevent water waste, the platform was built with irrigation and a water drainage system. The extra water
will be collected and sent to the water storage using the water drain system. Additionally, a UV lighting
system was built to aid plant growth in the lack of direct sunshine.
A. Experimental Design
Fig. 6 Different Data Mining Techniques
This study uses data mining to classify planting sustainability using abiotic and edaphic data gathered from the
indoor or greenhouse facility. Hidden information from databases is transformed into comprehensible
structures using a computational process called data mining. To find patterns in massive amounts of data, it
uses artificial intelligence, machine learning, statistics, and database systems. Predictive data mining is the
most popular type with direct business applications because prediction is the end goal[28].
There are several data mining techniques, but the primary techniques are the Classification, the Clustering, the
Association rule mining and Regression analysis. Fig. 6 displays the various data mining methods now in use.
This study focused on the Classification Techniques since the study is about predicting the planting
sustainability based on data collected from the prototype. To determine which algorithm would provide the
best fit and the highest accuracy, classification algorithms including Random Forest, Naive Bayes, and KNN
were tested.
1 ) Study Area:
The study was conducted in Novaliches, District 5, Quezon City where the improvised vertical farming
platform will be implemented. This is to support the Urban Farming Project of District 5, Quezon City,
Philippines. Barangay Kaligayahan, Quezon City is the project's beneficiary and recipient of the developed
prototype The prototype was tested in Bagong Silang. Caloocan City under the assistance of Mr. Eduardo
Torrest Jr., an entrepreneur and a lettuce grower.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1000
www.rsisinternational.org
Fig. 7 District 5, Quezon City[29]
2 ) Collect:
The study's primary data was taken from the prototype’s collected data. Data collected are presented to
agriculturists and experts to be given proper assessment and to identify the planting sustainability based on
values of different parameters. All data uploaded to the web server that was evaluated by experts were used as
datasets and were utilized in designing a predictive model.
3 ) Prepare Data:
During this phase, all data collected from the database were processed in order to transform raw data into a
useful and efficient format. The process involves removing missing, inconsistent, or irrelevant data, including
duplicate records, missing values, and outliers. The data in .csv format will be used data mining, with 70%
being training data and 30% being test data, ensuring a suitable format for the task.
4 ) Train Model:
The training instruments and provided training data are described in this section. In order to visualize data, this
study uses Orange Visual programming. The 70% of the data was utilized as training data in creating a model
that will give the highest accuracy in classifying the planting sustainability.
5 ) Test Data:
After the model was created, 30% of the dataset was used to test the model. Different models were tested to
determine which of these models employs different classification algorithms. The dataset is converted to a.csv
file and used machine learning models including KNN, Random Forest, and Naive Bayes. Evaluation results
are provided by the confusion matrix.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1001
www.rsisinternational.org
6 ) Classification and Validation:
The researcher assessed the performance, efficacy, and accuracy of the classifier model's predictions using
various measures. The reliability and validity of the data were determined through tests with known planting
sustainability parameter values. The accuracy of the generated sample was evaluated by comparing the
observed accuracy with an expected accuracy rate based on the Confusion Matrix[30]. Precision, Recall, and
F-1 Score are taken into account and determined as performance indicators. Table II displays Cohen's Kappa,
which is needed to properly interpret the performance of each classifier.
Fig. 8 Confusion Matrix [30]
Table II Magnitude of Kappa as Illustrated by Landis and Koch[31]
Kappa (K)
< 0.00
0.00 0.20
0.21 0.40
0.41 0.60
0.61 0.80
0.81 1.00
RESULT AND DISCUSSION
B. The Developed Prototype
The developed prototype is thoroughly discussed in this section. The prototype was created by integrating
various pieces of hardware, software, and data mining theory to make it function in accordance with its design
and specifications.
The prototype's IoT gateway is shown in Fig. 9. The Internet of Things gateway, sometimes called an IoT
gateway, acts as a link between WSN’s and the cloud infrastructure. IoT gateway serves as a primary hub for
sending to the cloud of abiotic and edaphic measurements. Arduino Wemos D1, relays, RFM95 LoRa
Transceiver, casing, and antenna are the critical parts to build the IoT gateway.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1002
www.rsisinternational.org
Fig. 9 The Developed IoT Gateway
Fig. 10 The Wireless Sensor Node
Figure 10, on the other hand, shows a WSN’s prototype that has been built and is in charge of reading
greenhouse climatic and soil parameters and transmitting all of the collected data to the IoT gateway via radio
frequency. The primary parts of the WSN are the Arduino Mega, RFM95 LoRa Transceiver, DHT11, NPK
sensor, LDR, soil moisture sensor, pH sensor, and TTL to RS485 module.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1003
www.rsisinternational.org
C. Mobile Application
Fig. 11 Smart Space Argi Mobile Application
The IoT gateway pushes data to the Blynk IoT Platform enabling the Android Application to display all the
collected data in addition to delivering the data into the cloud. By providing the most recent readings of
climatic and edaphic parameters, the mobile application keeps the end-user or farmer informed about the state
of his or her greenhouse facilities. The mobile application also enables remote control of the ventilation and
irrigation system for the user. The researcher employs widgets, including pushbuttons for remote controls and
gauges and graphs to display data. The programs use tabs titled Climatic, Soil, and Controls to organize the
presentation of data and controls.
D. Web Interface
Fig. 12 Smart Space Web Interface
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1004
www.rsisinternational.org
The end user can view all of this information through the system's web interface once the IoT gateway uploads
the greenhouse data that has been collected. A web interface was developed to enable end-users gain access to
green house’s edaphic and climatic condition remotely using any browsers and internet connection. The web
interface was hosted by CloudFare.com. Fig. 12 shows the screen shot of the web interface.
E. Comparative Analysis of Algorithm
Three Classification Algorithms were used in the comparative analysis, the KNN, Random Forest, and Naïve
Bayes. The researcher employs Orange Visual Programming to perform the trade-off.
The dataset has 4,438 occurrences and was divided into training and testing subsets. The isSustainable column
in the dataset comprises categorical data with the role feature "target." The training set is used to train the
classification algorithm, while the testing set is used to evaluate its effectiveness. 70% of the dataset goes
toward testing, and 30% goes toward training.
TABLE III Evaluation Results of the Algorithm
Model
AUC
CA
F1
Precision
Recall
Random Forest
0.995
0.995
0.995
0.995
0.995
Naïve Bayes
0.996
0.894
0.894
0.898
0.894
KNN
0.478
0.495
0.493
0.491
0.495
Table III shows the result of the comparative analysis in which Random Forest registers the highest AUC, CA,
F1, Precision and Recall with 0.995, 0. 995, 0. 995, 0. 995, and 0. 995 respectively. Second is Naïve Bayes
with 0.956 AUC, CA of 0.894, F1 of 0.894, Precision of 0.898, and Recall of 0.894. The last among the four
algorithms is the KNN with CA of 0.478, AUC of 0.495, F1 of 0.493, Precision of 0.491, and Recall of 0.495.
To avoid over fitting, the evaluation of the algorithms was performed using 10-fold cross validation.
Table IV Observed Accuracy, Expected Accuracy, and Kappa Values of Classifiers
Model
Observed Accuracy
Expected Accuracy
Kappa
Strength of Agreement
Random Forest
0.99512
0.50608
0.9901
Almost Perfect
Naïve Bayes
0.89410
0.50081
0.7879
Substantial
KNN
0.49531
0.50929
-0.0285
Poor
Table IV presented the Observed Accuracy, Expected Accuracy, Kappa Values and the Strength of Agreement
of the three classifiers used by the researcher. Random Forest registers the highest value of Observed Accuracy
of 0.99512, Expected Accuracy of 0.50608, Kappa of 0.9901, and “Almost Perfect in terms of Strength of
Agreement. Naïve Bayes has 0.89410 for Observed Accuracy, 0.50081 for Expected Accuracy, 0.7879 for
Kappa, and Substantial for the Strength of Agreement. KNN has 0.49531 for Observed Accuracy, 0.50929 for
Expected Accuracy, -0.0285 for Kappa, and Poor for the Strength of Agreement.
The result presented in Table IV illustrates that Random Forest registers the highest Kappa with an
interpretation of “Almost Perfect” in the strength of agreement.
F. System Evaluation
The user, agriculturist, and information technology experts evaluated the constructed prototype using the
ISO25010 standard as part of the prototyping methodology. The system was evaluated based on Functional
Suitability, Performance Efficiency, Compatibility, Usability, Reliability, Security, Maintainability, and
Portability. Three (3) IT industry specialists, one (1) agriculturalist, and one (1) lettuce farmer all took part in
the test. The evaluation's findings are shown in Table V.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1005
www.rsisinternational.org
Table V User Evaluation Results
Factor
Weighted Mean
Interpretation
Rank
Functional Suitability
4.35
Agree(A)
2
Performance Efficiency
4.33
Agree(A)
3
Compatibility
4.07
Agree(A)
6
Usability
4.36
Agree(A)
1
Reliability
4.25
Agree(A)
4
Security
4.10
Agree(A)
5
Maintainability
3.93
Agree(A)
7
Portability
4.10
Agree(A)
5
Over-all Mean
4.19
Agree(A)
Table V presents the result of the user evaluation for each criterion. The constructed system displays a high
level of usability as it received the highest ranking among criteria based on the computed weighted mean of
4.35, with verbal interpretation of "Agree (A)." Next came Functional Suitability, which received a weighted
mean of 4.25 and verbal interpretation of "Agree (A)," followed by Performance Efficiency, which received a
computed weighted mean of 4.33 and was verbally interpreted as "Agree (A)." Reliability has 4.25 weighted
mean and a verbal interpretation of “Agree (A).” Security and Portability shares the same rank with 4.10
weighted mean and a verbal interpretation of “Agree (A).” The Compatibility is ranked sixth with a computed
weighted mean of 4.07 and a verbal interpretation of "Agree (A)." Lastly, the Maintainability has the lowest
weighted mean of 3.93 and verbal interpretation of "Agree (A)". The developed prototype has done well in
terms of a number of quality parameters, as evidenced by the overall weighted mean of 4.19, with a verbal
interpretation of "Agree (A)". The outcome indicates that the system is of excellent quality, excelling in a
number of areas including functionality, dependability, usefulness, efficiency, maintainability, and portability.
It shows strong performance across several metrics, proving that it satisfies or even exceeds the standards
outlined in the ISO 25010 quality model.
CONCLUSION
An IoT gateway and a wireless sensor node (WSN) comprise the project's two components. An Arduino Mega
connects to the sensors and peripherals that the WSN is equipped with to collect data from the surroundings.
For long-distance, high-data-rate, error-prone communication, the WSN makes use of components such the
RFM95 LoRa Transceiver, DHT11, RS485 Soil NPK Integrated Sensor, Light Dependent Resistor, Soil
Moisture Sensor, pH Sensor, and Max485 TTL to RS-485 module. The IoT gateway offers connectivity,
protocol translation, data filtering, and local processing capabilities as it transmits the collected data to the
cloud platform. The IoT gateway utilizes the WeMos- D1 R2 WiFi ESP8266 Development Board and contains
a LoRa transmitter for data reception.Before being sent to the cloud platform, which requires internet
connectivity, all data received from the WSN is organized in JSON format.
The wireless sensor node, IoT gateway, web interface, and mobile application are the four parts that make up
the overall system. Utilizing LoRa technology, all captured data is transmitted over radio frequency into the
IoT gateway, enabling the system to consume less internet. Once data has been received by the IoT gateway
over the internet, it will be uploaded and stored in Firebase, a backend cloud computing service. The
constructed web interface allows the end-user to view and access all of these records. Only authorized users
can use the web interface, which is password and username protected. The web component can provide the
most recent data gathered from the greenhouse along with a projection of how long the planting will last.
Additionally, graphs are used to provide end users with a simple way to see the historical values of various
abiotic elements as well as the soil moisture level and nutrients from the wireless sensor nodes. Along with the
web components, a mobile application was developed to improve the system's usability and accessibility and
allow end users more access to the greenhouse. The mobile application shows the most recent sensor readings
from the gadget. The irrigation system and air-cooling system can also be managed by the farmer or end user
using the application.
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1006
www.rsisinternational.org
Random Forest, Nave Bayes, and KNN are the three classification methods that the researcher compares using
Orange Visual Programming. Based on the result, Random Forest Algorithm exhibits the highest level of
accuracy based on the computed Kappa of 0.9901, with an “Almost Perfectstrength of agreement. Random
Forest algorithm was implemented into the system’s web interface using a machine learning PHP library,
Rubix ML. All collected data is uploaded to the cloud and kept in the Firebase database using the constructed
wireless sensor node and IoT gateway. Based on the abiotic variables and soil nutrients found in the
greenhouse or indoor farm, the results have been classified and marked as "Sustainable" or "Not Sustainable"
in the database. The system can also measure soil moisture using a soil moisture sensor integrated into Arduino
Mega that triggers the sprinkler system when the water level drops at the threshold point or as the crop needs.
The overall weighted mean of 4.19, with a verbal interpretation of "Agree (A)", shows that the developed
prototype performed well in terms of quality characteristics stipulated in ISO25010. The result shows the
system demonstrated high quality, excelling in a variety of factors such as functionality, dependability,
usefulness, efficiency, maintainability, and portability. The results show that the system meets or exceeds the
requirements specified in the ISO 25010 quality model.
ACKNOWLEDGEMENT
The authors would like to extend special thanks to the employees of Philippine Department of Agriculture for
providing relevant data. Colegio de Sta. Teresa de Avila and Commission on Higher Education for the
financial support. We would also like to thank Dr. Jennifer Carpio, Dr. Albert Vinluan, Dr. Dennis Gonzales
and Luisito L. Lacatan for the guidance and support to this study.
REFERENCES
1. G. J. Perez, M. Macapagal, R. Olivares, E. M. Macapagal, and J. C. Comiso, “Forecasting and
monitoring agricultural drought in the Philippines, Int. Arch. Photogramm. Remote Sens. Spat. Inf.
Sci. - ISPRS Arch., vol. 41, no. July, pp. 12631269, 2016, doi: 10.5194/isprsarchives-XLI-B8-1263-
2016.
2. S. Farming, “Smart Farming,” no. 113, pp. 117, 2018, doi: 10.3390/s18114051.
3. M. Chetan Dwarkani, R. Ganesh Ram, S. Jagannathan, and R. Priyatharshini, Smart farming system
using sensors for agricultural task automation,” Proc. - 2015 IEEE Int. Conf. Technol. Innov. ICT
Agric. Rural Dev. TIAR 2015, no. Tiar, pp. 4953, 2015, doi: 10.1109/TIAR.2015.7358530.
4. E. O. Brown, R. V Ebora, and F. L. C. Decena, The Current State , Challenges and Plans for
Philippine Agriculture 1 Philippine Council for Agriculture , Aquatic and Natural Resources Research
This paper provided brief discussion of the current state of Philippine agriculture , the challenges being
fa,” vol. 2017, no. Table 1, pp. 116.
5. C. F. Habito and R. M. Briones, Philippine agriculture over the years: performance, policies and
pitfalls,Agriculture, p. 38, 2005.
6. N. Kaewmard and S. Saiyod, “Sensor data collection and irrigation control on vegetable crop using
smart phone and wireless sensor networks for smart farm, ICWiSe 2014 - 2014 IEEE Conf. Wirel.
Sensors, pp. 106112, 2014, doi: 10.1109/ICWISE.2014.7042670.
7. Louise Maureen Simeon, Philippines secures $2.43 million funding to build smart greenhouses |
Philstar.com,” The Philippine Star, 2018. https://www.philstar.com/business/agriculture/2018/ 07/ 08/
1831455/philippines-secures-243-million-funding-build-smart-greenhouses (accessed May 25, 2019).
8. H. Bach and W. Mauser, “Sustainable Agriculture and Smart Farming, Earth Obs. Open Sci. Innov.,
pp. 261269, 2018, doi: 10.1007/978-3-319-65633-5_12.
9. A. . Fallis, Managing water and fertilizer for sustainable agricultural intensification, vol. 53, no. 9.
2013.
10. C. Yoon, M. Huh, S. G. Kang, J. Park, and C. Lee, Implement smart farm with IoT technology,” Int.
Conf. Adv. Commun. Technol. ICACT, vol. 2018-Febru, pp. 749752, 2018, doi:
10.23919/ICACT.2018.8323908.
11. J. Pitakphongmetha, N. Boonnam, S. Wongkoon, and T. Horanont, “Internet of Things for Planting in
Smart Farm Hydroponics Style,” 2016.
12. S. Choi et al., “Design and implementation of a connected farm for smart farming system FIESTA-IoT
INTERNATIONAL JOURNAL OF RESEARCH AND SCIENTIFIC INNOVATION (IJRSI)
ISSN No. 2321-2705 | DOI: 10.51244/IJRSI |Volume XII Issue X October 2025
Page 1007
www.rsisinternational.org
View project Design and Implementation of a Connected Farm for Smart Farming System,pp. 14,
2015, doi: 10.1109/ICSENS.2015.7370624.
13. H. McNairn, A. Merzouki, A. Pacheco, and J. Fitzmaurice, Monitoring soil moisture to support risk
reduction for the agriculture sector using RADARSAT-2,” IEEE J. Sel. Top. Appl. Earth Obs. Remote
Sens., vol. 5, no. 3, pp. 824834, 2012, doi: 10.1109/JSTARS.2012.2192416.
14. Y. O. Sayad, H. Mousannif, and M. Le Page, “Crop management using Big Data,” Proc. 2015 Int. Conf.
Cloud Comput. Technol. Appl. CloudTech 2015, 2015, doi: 10.1109/CloudTech.2015.7337003.
15. R. O. Slayter, The influence of progressive increases in total soil moisture stress on transpiration,
growth, and internal water relationships of plants,” Aust. J. Biol. Sci., vol. 10, no. Meyer 1945, pp.
320336, 1957.
16. M. K. M. Nasution, M. Elveny, and R. Syah, “Research methodology for computer science,” AIP Conf.
Proc., vol. 2714, pp. 14, 2023, doi: 10.1063/5.0134511.
17. S. Dutta, A. Mitra, S. Chatterjee, S. Lodh, and S. Mukherjee, “A sensor-based approach to monitor a
specific plant sustainable environment with additional automated rehydration module, 2017 4th Int.
Conf. Opto-Electronics Appl. Opt. Optronix 2017, vol. 2018-Janua, pp. 15, 2018, doi:
10.1109/OPTRONIX.2017.8349986.
18. P. R. Harshani, T. Umamaheswari, R. Tharani, S. Rajalakshmi, and J. Dharani, “Effective Crop
Productivity and Nutrient Level Monitoring in Agriculture Soil Using IOT,ICSNS 2018 - Proc. IEEE
Int. Conf. Soft-Computing Netw. Secur., no. Cl, pp. 110, 2018, doi: 10.1109/ICSNS.2018.8573674.
19. G. Kaur, P. Upadhyaya, and P. Chawla, Comparative analysis of IoT-based controlled environment
and uncontrolled environment plant growth monitoring system for hydroponic indoor vertical farm,
Environ. Res., vol. 222, no. February, p. 115313, 2023, doi: 10.1016/j.envres.2023.115313.
20. S. Bhardwaj, T. Ozcelebi, J. J. Lukkien, and K. M. Lee, “Smart space concepts, properties and
architectures,” IEEE Access, vol. 6, no. c, pp. 7008870112, 2018, doi:
10.1109/ACCESS.2018.2880794.
21. P. Dias, J. Rodrigues, A. Aguiar, and G. David, Planning and managing data for Smart Cities: An
application profile for the UrbanSense project,2018 IEEE Int. Smart Cities Conf. ISC2 2018, 2019,
doi: 10.1109/ISC2.2018.8656835.
22. A. Vasilev, E. Krylov, A. Subbotkin, and I. Paramonov, Event recording system for smart space
applications,” Conf. Open Innov. Assoc. Fruct, vol. 2012-Novem, pp. 147153, 2017, doi:
10.23919/FRUCT.2012.8122099.
23. S. Chae, Y. Yang, J. Byun, and T. Han, “Personal Smart Space : IoT based User recognition and
Device control,” pp. 5–6, 2016, doi: 10.1109/ICSC.2016.45.
24. P. Mtshali and F. Khubisa, “A smart home appliance control system for physically disabled people,”
2019 Conf. Inf. Commun. Technol. Soc. ICTAS 2019, pp. 15, 2019, doi:
10.1109/ICTAS.2019.8703637.
25. T. Wiangtong, “IoT-based Versatile Platform for Precision Farming,2018 18th Int. Symp. Commun.
Inf. Technol., no. Iscit 2018, pp. 438441, 2020.
26. “Prototype Model in Software Engineering.” https://www.guru99.com/software-engineering-
prototyping-model.html (accessed Sep. 25, 2023).
27. Q. Requirements, “INTERNATIONAL STANDARD ISO / IEC,” vol. 2011, 2011.
28. M. Paul, S. K. Vishwakarma, and A. Verma, “Analysis of Soil Behaviour and Prediction of Crop Yield
Using Data Mining Approach,” Proc. - 2015 Int. Conf. Comput. Intell. Commun. Networks, CICN
2015, pp. 766771, 2016, doi: 10.1109/CICN.2015.156.
29. “District 5 - Quezon City Great, Green, Growing.” http://quezoncity.weebly.com/district-5.html
(accessed Sep. 01, 2023).
30. S. Bittrich, M. Kaden, C. Leberecht, F. Kaiser, T. Villmann, and D. Labudde, Application of an
interpretable classification model on Early Folding Residues during protein folding 08 Information and
Computing Sciences 0801 Artificial Intelligence and Image Processing,” BioData Min., vol. 12, no. 1,
pp. 117, 2019, doi: 10.1186/s13040-018-0188-2.
31. J. R. Landis and G. G. Koch, “The Measurement of Observer Agreement for Categorical Data
Published by : International Biometric Society Stable URL : http://www.jstor.org/stable/2529310,”
Biometrics, vol. 33, no. 1, pp. 159174, 1977.