1. data to make professional decisions. Soon using

1.     
Abstract

Farming is largely a challenging and
cumbersome profession. The unstable and volatile commodities market squeezes
the life out of a farmer who already is experiencing unprecedented scarcity of
water on top of ever-rising operational costs. 
Farmers are constantly subjected to restrictive regulations on
irrigation, pesticide use and fertilizer application, which leads them to
explore and find new ways to boost the agricultural yield. Fortunately, a huge
amount of data is available on modern farms ranging from yield monitors to
infrared imaging, but the sad state of affairs that agricultural profession is
ages behind other industries in utilizing data to make professional decisions.
Soon using data to optimize decision making will no longer be a novelty, but an
essential practice to stay afloat in business. This paper discusses different
decision making algorithms. The challenge identifying predictive abilities
using promising methods with a small dataset to predict agricultural yields and
the performance and the characteristics of different machine learning
algorithms have been discussed.

We Will Write a Custom Essay Specifically
For You For Only $13.90/page!


order now

2.     
Introduction

 

The world population is constantly in an
upward omentum in the background of unpredictable climatic changes. Farmers are
kept on their toes as they face the conundrum of making difficult decisions as
to how to remain productive and sustainable in the ever changing climatic and
economic conditions. Using computers could help farmer ward of these
challenges; they can use computers in assessing the fertility of the soil
nutrient, what fertilizers and pesticides best suit their land, etc. This could
help them in attaining greater crop productivity if the conditions are suitable
and decrease their loss due if the conditions do not favour them (Niketa Gandhi et al, 2016). Crop
growth, quality and yield are highly dependent upon weather and environmental
factors such as seasonal temperature and precipitation variations, day-to-day
temperature ranges and, water cycles between soil and atmosphere. As in every
vocation, less productivity in agriculture translates to increased price of the
produce in the market. In this backdrop, machine learning algorithms will help
the farmers to double the crop production. Better crop predictions help farmers
improve their nitrogen management to meet the demands of the new crop and mill
managers could better plan the mill’s labor requirements and maintenance
scheduling activities, and marketers can more confidently manage the forward
sale and storage of the crop. Hence, accurate yield forecasts can improve
industry sustainability by delivering better environmental and economic
outcomes. The predictor variables include variables based on indices for
simulated biomass, previous yields, local climate data consisting of rainfall,
radiation, and maximum and minimum temperature. Big data technologies increase
industry guidance on key industry decisions that affect sustainable
agricultural systems and contribute a partial solution to food shortages.

This paper has been structured in the
following manner: Section 2 explains the Architecture of Smart Agriculture
Management System (SAMS).  Section 3
explains the importance of Predictive Analytics in Precision Agriculture.  Section 4 explains the Survey on Crop yield
Prediction. Section 5 explains the inference and conclusion of this study.

3.     
Architecture
of Smart Agriculture Management System (SAMS)

                                                                 

Sensors

Sensors

Sensors

 
 
Cloud Server
Data Analytics

Actuators

User

Gateway

Gateway

Gateway

Layer1

Layer2

Layer3

           

 

 

 

 

 

 

 

 

 

 

Figure 1: Smart Agriculture Management System

 

Figure 1 shows the architecture of Smart Agriculture
management Systems, which has three layers. The first layer comprises sensors
and gateways where the physical environment changes are monitored and sent to a
cloud server to perform analytics through the gateway. The second layer
comprises a cloud server in which Analytics Algorithm can be run and smart
decisions be made. Then the decision would be sent to the third layer, where
the user or actuators will respond to the decision. In this survey we are
focusing on algorithms involved in Data Analytics of Layer 2.

In the concept of IoT, the server should
be intelligent enough to make decisions appropriately. The sensor monitors the
soil moisture, leaf wetness, temperature, and humidity level in the environment
and sends the data to the server through the gateway in which it performs the
analytics and then the server sends the recommendations to the farmer’s mobile
phone on which the actions of the farmers can be based. (Raheela, 2016).

4.     
Predictive
Analytics in Precision Agriculture

Because
of increasing demand of decision support systems, Precision Agriculture can be
used as an effective tool.  The services
that can be obtained using Precision Agriculture are information services,
traceability systems, precision irrigation, monitoring, controlling and
management of the field (Shailaja
Patil,2016).  With the help of
Predictive Analytics, we can get realtime data on climate, soil and air
quality, planting, crop maturity, equipment and labour costs and availability.
These data will help us in making an ideal decisions regarding a sustainable
agricultural development plan.  The goal
is to get better understanding of the land, weather, climate and planting,
which can help in prediction of events with greater accuracy which in turn can
easily translate to sustainable agriculture (Khushboo Babaria,2015). We have 4 different steps in this
process;  (1) fresh data gathering and
cleaning; (2) renovating the cleaned data into a desired format that can be
used by the machine learning method; (3) generating a predictive model
(training) using the renovated data; (4) reporting predictions to the user
based on the previously created predictive model. The learning model applies
the following areas in agriculture domain:

1.      Prediction
of Crop Yield

2.      Plant
Diseases Detection and Classification

3.      Management
of Fertilizers and Pesticides

4.      Ranking
and Categorizing  of Agriculture Products

5.      Supervision
of Soil Fertility

5.      Literature Survey

This
survey will give a brief narration about classification algorithm in machine
learning, which can help in building a model to predict crop yield. These
predictions will help the farmer to produce good yield and maintain the
sustainability of soil. A few classification algorithms are discussed in this
survey in the perspectives of crop prediction; they are Support Vector Machine,
Naïve Bayes, Neural Network and Random forest.