The Research on the Application of Plant Identification and Mobile

Learning APP based on Expert System

Cixiao Wang

Graduate School of Education, Peking University, No.5 Yiheyuan Road Haidian District, Beijing, P.R. China

Keywords: Expert System, Plant Identification, Mobile Learning, Application Research, Outdoor Teaching.

Abstract: As the development of mobile internet and the improvement of software and hardware facilities, expert

system expands its field in mobile applications. The identification and learning of plants is an important part

of biology in middle school. Outdoor experiential teaching is an effective way to improve the students'

learning effect. Based on the production rules expert systems, this study developed plant identification and

learning mobile app based on Android platform. This study applied the app into the outdoor mobile learning

and used the education experiment and classroom observation method to do research by the use of software,

attitude of using software, learning satisfaction learning and attitude towards science four dimensions. And

the analysis of the difference between the population statistics and the application of mobile learning

experience, verify that the system has a positive effect on the students' learning attitude and the degree of

satisfaction with the outdoor experiential teaching.

1 INTRODUCTION

The expert system is a heuristic program system

which can use expert knowledge to do empirical

reasoning. It contains a large number of expert level

domain knowledge, which can simulate the

reasoning process of a human expert to solve

problems (Jia, 2009). In 1965, Fei Baum in Stanford

University and chemist Lud Begg developed the first

expert system DENDRAL (Shi, 2011). Since the

1980s, the expert system is gradually applied to

various fields with the development of science and

technology (Pandit, 2013). In the field of education,

the application of expert system embodied

intelligent teaching system, intelligent question

answering expert system, intelligent decision

system, intelligent test paper evaluation system, etc..

(Golumbic et al., 1986; Xu and Jiang, 2011; Zhu,

2012) With the popularity of smartphones and the

development of mobile Internet, mobile learning has

become a new field of expert system in educational

application (Fu and Li, 2010).

Biology learning in middle school is the basic

stage, and knowledge about plant is one of its

components. The plant identification is more

complex than animal identification, which is the key

to biology learning, need more practice to carry out

experiential learning in a real environment (Hou et

al., 2012). The plant recognition and learning mobile

application based on expert system can make

students experiencing the process of plant

knowledge learning outdoors, by which students can

learn knowledge of plant genera and characteristics

flexible and fully.

2 RESEARCH STATES

2.1 Limitation of Plant Identification

With the development of science, now the

identification of plant can be roughly divided into

three categories (Chen et al., 2014), artificial

identification method, assisted artificial

identification method and automatic identification

method. Artificial identification method refers to the

plant characteristics of knowledge investigation

form of learning, such as flora and botany etc..

Corresponding to the actual life, plant identification

is divided into a visual method, smell method,

somatosensory method. The method requires experts

to master a wide variety of plant characteristics

knowledge. Experts in the field can quickly identify

plants through this method, while middle school

students are not competent. Assisted artificial

identification method id using the existing data in

332

Wang, C.

The Research on the Application of Plant Identiﬁcation and Mobile Learning APP based on Expert System.

DOI: 10.5220/0006313103320339

In Proceedings of the 9th International Conference on Computer Supported Education (CSEDU 2017) - Volume 2, pages 332-339

ISBN: 978-989-758-240-0

physical or chemical methods to help people identify,

such as simple tools, microscopy, spectroscopy,

thermal spectroscopy and other high-precision

methods. However, this method is not suitable for

ordinary biology classroom or outdoor experiential

teaching. Automatic identification method is using

automatic identification system based on computer

vision to observe leaf characteristics. Computer

vision technology can automatic complete plant leaf

image processing and feature extraction and

classification of plants. While this method is time-

consuming, can not provide instant feedback on

mobile learning.

2.2 Relevant Research

Mobile applications contribute to project-based

learning, problem-based learning, and other

integrated practical activities, to develop students'

ability to communicate, solve problems, innovation

and innovation ability.

Huang(Huang et al., 2010)

developed a Mobile Plant Learning System (MPLS)

based on the pad, which provides outdoor

experience to recognize plant and learn botany

knowledge in the primary school curriculum. MPLS

belongs to the framework based expert system, in

which stored a large number of plant leaf

characteristics and detailed examples of information.

Through the comparison between pre-test and post-

test in the experimental group, it was found that

through MPLS learning, students' ability of plant

recognition was improved obviously, and the

outdoor learning method was more popular.

Mobile applications based on interactive concept

maps are also applied in middle school biology

learning. (Hwang et al., 2011)Research shows that

instant feedback of mobile application learning

method is conducive to improve students' interest in

learning and outdoor biological science teaching

effect. The practical teaching system of campus

plant scene teaching is designed (Xu et al., 2015),

which includes pre-class learning and outdoor

experiential learning in class and teaching feedback

underclass.

Compared with the traditional classroom

knowledge teaching, outdoor experiential learning is

more helpful to improve students' interest in

scientific knowledge and knowledge of plant

knowledge. The mobile application expert system

can promote the application of outdoor mobile plant

identification and learning of middle school students

not only need simple and easy to operate, plant

information database based on large, there should be

immediate feedback operation, help learners to

quickly complete plant identification, and learn more

knowledge about plant characteristics.

3 APPLICATION

DEVELOPMENT

3.1 Expert System

The expert system based on rules also called the

generative rules system, there are many examples of

successful and simple and flexible development

tools, can directly imitate human psychological

process, and use a series of rules to express expert

knowledge (Zhang et al., 2010). This study

established a plant facts database of non-attribute

rules, including the fact of plants, that is attribute

value of the attribute refers to a "yes" or "no", which

is a series statement of IF and THEN. Figure 1 is a

simplified structure of the expert system which was

designed for this research. Through the display of

obvious plants, the system has many aspects

describing the entries and simple image schematic.

The learner can answer the question when observing

plants, then the reasoning machine and the

interpreter in the knowledge database shows the next

question refers to feedback from the learner. Until

the correct reasoning to plant so far, all the

information about the right plant will be shown in

the result page on the app (including plant number,

picture, name, alias, characteristics, distribution and

use value) for the learners.

Figure 1: Simplified architecture of plant recognition and

learning expert system.

3.2 System Inference and Interpreter

In this study, the inference engine design of

production rule expert system can enable learners to

identify plants through a variety of plant

characteristics. The interpreter is the plant fact

information, which combines the operation of the

corresponding machine. This study designed the

intermediate facts to simplify the inference process

of plant identification, that is, let learners judge plant

The Research on the Application of Plant Identiﬁcation and Mobile Learning APP based on Expert System

333

genera (herbaceous, woody and Fujimoto) first.

Figure 2

structure for plant mainly adopts two query

tree data, the first access node, if the user chooses

"yes", then access the left subtree, or visit the right

subtree, until no node is given, the plants need to

query. The asterisk represents a picture prompt.

Each question makes learners observe plant

characteristics and make decisions, greatly the

promoting learner's participation in learning and

improving the frequency of learners’ observing of

plant characteristics.

Figure 2: Sketch data structure of plant recognition and

learning mobile applications.

Note: When the question number is greater than 1000, indicating

that the node of binary tree has come to an end, which means the

user can get the answer. The question marked with an asterisk has

picture prompts.

3.3 Plant Knowledge Base

Plant knowledge base formed in this study included

a total of 204 species those were common in the

northern campus in China and have obvious

characteristics of plants, including 78 species of

woody herb, 99 species, 27 species of vines.

According to "Chinese flora" (editorial board,

Chinese Academy of Sciences Chinese flora 2004),

this study wrote the two retrieval disambiguation

table (shown in Table1).

Table 1: Plant retrieval tables (part).

1. Plants are woody or herbaceous.

2. Plants are woody plants.

3, trees, plants with a single trunk

4, leaf needle shape, scale shape, thorn

shape, evergreen

5, leaves all scale flake, interactive pairs,

twigs flat, arranged in a plane...........................

Platycladus orientalis

5, needle-like, rod

6, leaf type two, for scaly leaves or

barbed leaves, branchlets terete....................................

Sabina Chinensis

6, at least 2 gold leaf clusters

7, Ye Dansheng, spiral

arrangement

8, cone upright; leaf blade

back bar, two holes, spirally arranged; not a leaf...

Liaodong fir

8, pendulous leaves with a

prominent seat;

9, young branchlets hairy; leaves

subulate tip blunt; four white hole line........................

Picea Meyer

9, young branchlets smooth; tip

acuminate or acute; the stomatal band is not

obvious.............................. Picea wilsonii

The software knowledge database mainly includes

two tables: plant retrieval table (see Table 2) to find

the role of plant, provide answers; Plant information

table detailed information for the storage of plants,

which provide more learning content (including

plant number, picture, name, alias, characteristics,

distribution and use value) about the plant for

learners to learn.

Table 2: Design of plant retrieval table in knowledge base.

property has_pic question_id question yes no

Meaning

Whether

there is

picture

tips,

Picture

Question

Content

User

select

Yes,

roblem

number

User

select

"no",

question

number

data type Interger varchar varchar varchar varchar

3.4 Plant Identification and Learning

Process

Figure 3: Plant identification and learning process

(Magnolia).

CSEDU 2017 - 9th International Conference on Computer Supported Education

334

The application is developed by eclipse and Android

ADT, which can run in more than 2.3 versions of the

Android mobile platform. In the software testing

phase, and the average duration of the plant

identification process is 2 minutes.

Take the identification and learning of Magnolia

for example, when learners observe magnolia, and

choose wood for the entrance, and then answer a

series of questions by observing the features of

plants as shown in Figure 3 and Figure 4. Finally,

learners can obtain the graph and information of

magnolia (shown in Figure 5).

Figure 4: Question interface.

4 APPLICATION RESEARCH

4.1 Research Method

4.1.1 Education Experiment Design

This study conducted two education experiments on

plant identification and learning APP. The purpose

of Experiment 1 is to explore the use and attitude of

learners on the system based outdoor experiential

teaching. The purpose of Experiment 2 is to explore

the teaching extension of the system in different

stages of middle school.

Figure 5: Plant presentation interface.

Experiment 1 implemented in Beijing City,

Shijingshan District Yangzhuang middle school,

which is greening and rich in plant species.

Experimental design control group and experimental

group. The experimental steps of the experimental

group are as follows: (1) preview the relevant

knowledge of plant characteristics before class; (2)

use outdoor experience plant recognition and

learning in APP; (3) study and use. The

experimental steps of the control group were as

follows: (1) preview the knowledge of plant

characteristics before class; (2) learning knowledge

of plant recognition under the traditional classroom

teaching environment; (3) investigation of the

learning situation.

Experiment 2 is generalized inquiry experiment.

The experimental sites were selected from Taiyuan

forty-eighth middle school in Shanxi province. To

explore the use and attitude of the senior middle

school students, we choose the were senior high

school students as the experiment subjects. The

experiment steps are as follows: (1) preview the

plant characteristic knowledge before the experiment;

(2) use the APP to carry out outdoor experience

plant recognition and learning; (3) a survey of

learning and use.

In addition to the application of questionnaires in

educational experiments, the research method also

The Research on the Application of Plant Identiﬁcation and Mobile Learning APP based on Expert System

335

includes classroom observation and interview. To

observe the learning situation and attitude of the

learners participating in the outdoor experiential

plant recognition and learning, and the informal

interviews after class, the software usage is analyzed

comprehensively.

4.1.2 Research Tool

This study refers to the questionnaire used by Chu’s

(Chu et al., 2010) plant mobile learning applications,

and we combined with the cognitive characteristics

of mainland Chinese students, revised the scale of

each topic for translation and part of the topic of

expression. "Plant identification and learning mobile

APP usage questionnaire" is divided into scale part

and background information part.

Scale into two parts, a total of 25 questions (see

Table 3): plant identification mobile applications

using the first part of the survey students learning

activities, divided into three dimensions: "the use of

software" 4 questions, "the use of the software

attitude" 6 questions, "the use of the software to

learn after the learning satisfaction" 8 questions; the

second part is "to participate in the learning

activities, learning attitude of natural science",

namely the survey of students' attitude towards

natural science learning, a total of 7 questions. The

questionnaire uses the Likert five-point scoring

method, the option settings from "very agree" to

"very disagree", in turn, scored for 5, 4, 3, 2, 1. The

overall scale of Cronbach's Alpha coefficient is

0.894, the use of software, the use of the software of

attitude, learning methods and the satisfaction of

natural science learning attitude of Cronbach's Alpha

were 0.742, 0.793, 0.902, 0.907, indicating that the

research tools have higher reliability.

Table 3: Survey scale of plant identification and learning

mobile APP.

Construct Item

Use of software

This software is easy to use

I used a very short time to learn

how to use this software

The user interface of the software

is friendly

During the use, I did not

encounter technical problems

(such as click no response, etc.)

Attitude of using

software

I like to use this software to

learn, it provides an interesting

way to learn

I hope this way of learning can

be applied to other courses.

I will use this software to learn

I will recommend this software

to other students

Using this software makes

learning activities more

interesting

When using the software I can

seriously according to the tips of

each step to the next judgment

Satisfaction with

learning methods

after using the app

for learning,

This learning activity, let me

know more about the

characteristics of different plants

During my study, I have been

trying to observe the differences

between plants.

I learned to identify a plant by

observing the characteristics of

the flower, leaf, and fruit of the

plant.

Using software to learn is more

challenging and interesting than

traditional learning methods.

I learned new knowledge or new

discoveries

I try new ways of learning and

thinking

I learned how to identify a plant

I learned how to tell the

difference between plants

Learning attitude

towards natural

science after

participating in

learning activities

I am more interested in observing

and exploring plants.

I have more confidence in plant

learning

I am more interested in plant

learning

My observation of outdoor plants

has increased.

I hope to learn more about plants

I like to learn about the real

world through outdoor

observation.

4.2 Result Analysis

4.2.1 Analysis of Overall Effect

In Experiment 1, the research object is middle

school students from Beijing Shijingshan District

Yangzhuang junior high school, including students

from 1

class and 2

class in Grade three. The

students were volunteered to participate in this

experiment. 1

class is the experimental group, a

total of 15 people (7 boys, 8 girls), 2

class is the

control group, a total of 15 people (6 boys, 9 girls).

Experimental group students after the application of

plant identification fill in the questionnaire, the

control group only fill in the questionnaire

background information and scale fourth parts. A

CSEDU 2017 - 9th International Conference on Computer Supported Education

336

total of 30 questionnaires were issued, and 29 valid

questionnaires were recovered.

After the experimental group of learners using

the learning system, "the use of software", " attitude

of using software " and "learning satisfaction" scores

were in the middle level (4.883, 4.044, 3.733). It

shows that the system has a higher user experience

in software design, and the students in the

experimental group have higher satisfaction in using

the system. In the dimension of "learning attitude

towards science", the experimental group was filled

out after using the system, while the control group

did not use the system. After independent sample T

test, there was a significant difference between the

experimental group and the control group, that is,

the experimental group in the natural science

learning attitude, far higher than the control group.

This shows that the learning attitude and interest of

the natural sciences were greatly improved after the

outdoor activities of plant identification. The

purpose of this study, to improve the students'

interest in plant learning and plant characteristics

and knowledge come true.

Table 4: Difference between experimental group and

control group.

Dimension gender num

mean

value

standard

deviation

value

Learning

attitude

towards

natural science

experimental

group

15 3.990 0.360

7.619

0.0

00*

control

group

14 2.980 0.354

Note: * * * represents p<0.001.

4.2.2 Analysis of Teaching Extension

The subjects of Experiment 2 were students of grade

one in Taiyuan No. forty-eighth high school. A total

of 45 questionnaires were distributed, and a total of

35 valid questionnaires were collected, including 10

boys and 25 girls. The differences of experimental

results between students in experiment 1 and 2 were

analyzed. The differences in grade analysis results

are shown in Table 5, all dimensions are

significantly different. Senior high school learners of

the learning system, the use of software "," software

attitude "and" learning satisfaction "and" natural

science learning attitude scores were in the middle

level, and the scores were 4.671, 4.586, 4.668, 4.535.

The dimension of the use of the software is higher

than that of high school students in junior high

school, while high school students are higher than

junior high school students in the attitude of using

software, the satisfaction of learning methods, and

the attitude towards learning science dimensions.

Although the junior and senior high schools close in

grade, but the junior high school students will face

the senior high school entrance examination

(Biology is not included in the examination), high

school students have finished senior high school

entrance examination (while the college entrance

examination includes biology), and these factors on

learning application will have a certain impact.

Table 5: Difference between middle school students and

high school students.

Dimension grade num

mean

value

standard

deviation

t-value p

the use of

software

middle 15 4.883 0.229

2.431 0.019*

high 35 4.671 0.302

attitude of

using

software

middle 15 4.044 0.434

-3.501 0.001**

high 35 4.586 0.526

learning

satisfaction

middle 15 3.733 0.398

-7.554 0.000***

high 35 4.668 0.402

learning

attitude

towards

science

middle 15 3.990 0.360

-3.873 0.000***

high 35 4.535 0.489

Note: * represents P < 0.05, * * represents P < 0.01, * * *

represents p<0.001.

Overall, the system also has higher teaching

popularization value in senior high school, and it can

promote the senior high school students improving the

learning attitude of plant knowledge.

The development of

the educational application of relevant knowledge content

has a higher

feasibility and application value to help

middle school students and high school students to

conduct outdoor experiential learning.

4.2.3 Analysis of Differences in Application

The study used survey data of 50 learners to analysis

differences in demographic and mobile application

learning experience and other independent variables.

When to use the network to study there were no

significant differences in the use of software, the use

of software attitude, learning methods and learning

satisfaction attitude dimensions of natural science,

but with the delayed network learning time, each

dimension scores were increased, but the difference

is not large, indicating when starting study on the

use of significantly influence the recognition and

application of plants using network. The average

comparison shows that boys use software better than

girls, while girls are better in satisfaction of learning

and learning attitude towards natural science than

boys.

When to start using the network to learn in the

use of software, the attitude of using the software,

The Research on the Application of Plant Identiﬁcation and Mobile Learning APP based on Expert System

337

the satisfaction of learning methods, natural science

learning attitude dimensions were no significant

difference. With the delay of the beginning of using

the internet to learn, each dimension scores were

improved, but the difference is not large, which

indicating when to start to learn by the internet has

no significant effect on the identification and

learning of plant.

Whether has experience in using mobile

applications for learning before has a significant

difference in the use of software dimension, that is

the scores of students who have used the mobile

application to learn before were significantly higher

than those who haven’t. This shows that a certain

mobile application learning experience has a good

impact on the use of mobile applications, while there

are no significant differences in other dimensions.

Table 6: Difference in mobile application learning

experience.

Dimension

item

num

mean

value

standard

deviation

t-value

the use of

software

yes 37 4.777 0.287

1.537 0.048*

no 12 4.583 0.289

attitude of using

software

yes 37 4.356 0.593

-1.312 0.196

no 12 4.597 0.399

learning

satisfaction

yes 37 4.307 0.625

-1.887 0.069

no 12 4.604 0.412

learning attitude

towards science

yes 37 4.351 0.509

-0.442 0.661

no 12 4.429 0.578

Note: * represents P < 0.05, * * represents P < 0.01, * * *

represents p<0.001.

There are significant differences between "0-3h", "7-

15h" and "15h" group in the two dimensions of

learning method satisfaction and learning attitude

dimension of natural science. The longer the use of

mobile learning, the higher in the score in learning

satisfaction and learning attitude towards science

dimensions. There are similar rules in the use of

software and software attitude. It shows that if

students have a mobile applications learning

experience, the longer the use of mobile learning

applications, the better the use and learning results.

5 CONCLUSIONS

Combined with classroom observation and informal

interviews and data analysis of the above application

results, the following conclusions can be drawn:

(1) Using mobile learning applications to study

plant identification is popular with middle school

students. The use of software, the use of software

attitude, learning methods and learning satisfaction

attitude dimensions of natural science are higher

than traditional teaching methods, especially in the

use of software dimension;

(2) Using the plant identification of mobile

applications to learning knowledge outdoors can

make dimensions significantly higher than that of

students who haven’t in the natural sciences

students' learning attitude. The application of plant

identification significantly improves students in

science learning interest and attitude;

(3) The teaching extension of the system is quite

feasible. High school students in the

usage of software

dimensions are lower than the junior middle school

students. High school students in other dimensions is

higher than the junior middle school students;

(4) Gender had no significant effect on the use of

mobile learning in plant identification;

(5) When to start using the internet for learning,

there is no significant impact on the use of learning

in plant recognition mobile learning applications;

(6) Students who have used other mobile

applications for learning have scored significantly

higher on software usage dimensions than those who

did not use other mobile applications before learning;

(7) Have the experience of using mobile

applications to learn is helpful to use the app. The

longer the use, the use of mobile learning

applications and the better the learning effect.

6 DEFICIENCIES AND

PROSPECTS

Although plant learning in middle school is not the

emphasis part, the learners have a higher interest in

using the app to learning plant identification and

were enjoyable with the teaching methods. We can

see that outdoor experiential learning can bring

students a different experience from the traditional

classroom teaching. Through plant identification and

learning APP, learners can answer the system

questions given by the app, meanwhile, they can

observe plant characteristics, which can greatly

improve the learners' learning participation and the

understanding and mastery of plant knowledge. The

study shows that the app can improve students'

learning attitude towards natural science, stimulate

the learners' interest in learning related subjects such

as plants.

CSEDU 2017 - 9th International Conference on Computer Supported Education

338

This study makes a general analysis on software

usage and learning, but not for learners to plant-

related knowledge test to value learners' mastery of

the knowledge of plants. The further experimental

investigation is still needed. We should expand the

sample size at the same time, and add a degree of

data collection for students to master relevant

knowledge in the pre-test post-test. This study

designs an expert system based on production rules,

It consists of the knowledge base, inference engine,

and interpreter. However, the function is relatively

simple, as teaching interaction is limited to the

interaction between learners and teaching materials,

and is a lack of interactive teaching based on mobile

internet interaction and teacher-student interaction.

In addition to outdoor experience learning, we

should also provide plant information feedback, peer

exchanges, and cooperation, teacher feedback which

is equally important. The future research direction of

this study is to increase the communicative function

of learners, conducting peer collaborative inquiry

learning mode with outdoor experiential learning

under the guidance of teachers.

REFERENCES

Chen Q., Chen Y. F., Liu H., research status and Prospect

of 2014 plant identification methods. World forestry

research, 18-23.

Chu, H. C., Hwang, G. J., Tsai, C. C., Tseng, J. C. R.,

2010. A two-tier test approach to developing location-

aware mobile learning systems for natural science

courses. COMPUT EDUC 55, 1618-1627.

Fu G. S., Li T, mobile learning model based on expert

system in the 2010 3G era.

Golumbic, M. C., Markovich, M., Tsur, S., Schild, U. J.,

1986. A Knowledge-Based Expert System for Student

Advising. IEEE T EDUC e-29, 120-124.

Hou X. D, Zheng M, G. E. JW, 2012 how to arouse

students' interest in plant biology. Journal of biology,

105-107.

Huang, Y., Lin, Y., Cheng, S., 2010. Effectiveness of a

Mobile Plant Learning System in a science curriculum

in Taiwanese elementary education. COMPUT EDUC

54, 47-58.

Hwang, G. J., Wu, P. H., Ke, H. R., 2011. An interactive

concept map approach to supporting mobile learning

activities for natural science courses. COMPUT

EDUC 57, 2272-2280.

Jia J. Y., 2009 educational technology and artificial

intelligence. Jilin University press.

Pandit, M., 2013. Expert System-A Review Article .

International Journal of Engineering Sciences &

Research Technology 2.

Shi Z. Z., 2011 advanced artificial intelligence. Science

Press.

Xu T. Y., Jiang X. J., 2011. The application of artificial

intelligence based on expert system in the field of

education. Science and technology information, 31-55.

Xu Y. H, Xie C. P., Nan C. H, 2015 research on the

practice teaching system of identification and

identification of forest plants -- Taking campus plant

scene teaching as an example. Journal of biology, 106-

109.

Zhang Y. D, Wu L. N., development of 2010 expert

system. Computer engineering and application, 43-47.

Zhu L. C., the application and development trend of 2012

expert system in distance education. Journal of Hebei

Radio and TV University, 26-28.

The Research on the Application of Plant Identiﬁcation and Mobile Learning APP based on Expert System

339