Agriculture Productivity Performance of North-Western Zone in Tamilnadu Using Data Mining Techniques

 

C. Arul Kumar ¹, G. Manimannan^2*

¹Assistant Professor, Vivekanandha, College of Arts and Sciences for Women, Tiruchengode, Namakkal, Tamilnadu, INDIA.

²Assistant Professor, Madras Christian College, Tambaram, Chennai, Tamilnadu, INDIA.

 ^*Corresponding Address:

[email protected]

 

Abstract: Agriculture sector is the mainstay and backbone of the Indian economy. Agriculture plays a vital role in the development of a country. It contributes nearly fifteen percent of Gross Domestic Product (GDP) of India. Seventy percent of the population depends on agriculture for their livelihood. Over the past decade the agriculture production in districts of Tamil Nadu had faced increased yield in all the crops. However, agriculture productivity differs from region to region, which needs a detailed investigation. The main objective of this paper is to analyze the agriculture productivity of Major Crops in North-western zones of Tamil Nadu using Data Mining and Geographical Information System (GIS) techniques. The data was collected from secondary source of Department of Economics and Statistics, Tamilnadu during the period of 2003 to 2012. In this study we examined yield variations, spatial patterns and classification of fifteen major crops. The results achieved three different classifications and are labeled High Yield Districts (HYD), Medium Yield Districts (MYD) and Low Yield Districts (LYD) based on their mean yield of various crops.

Keywords: Agriculture Productivity, Productivity Regions, Crop Yield, Factor Analysis, k- mean Clustering Techniques, Geographical Information System and Multivariate Discriminant Analysis

 

Introduction

Agriculture is the back bone of our India. Agriculture and its allied sectors contribute nearly 15 percent of Gross Domestic Product (GDP) of India. Over 70 percent of the population is significantly dependent on agriculture for their livelihood. This sector provides 50 percent of the employment at the national level and it helping for the growth of the economy.The agriculture development is a requirement not only providing the food security of the growing population but also a number of industrial raw materials. In nutshell, agriculture plays an essential role in the process of economic development of the country. Agriculture continues to the major predominant sector of the Tamilnadu economy. The state has a total area of 1.3 lakh sq. km with a gross cropped area around 63 lakh hectares and the contribution to Gross Domestic Product and state domestic product, has higher compared with other sector. Agriculture development of a country or region is dependent on the production of crops. Several efforts have been taken to increase the production and productivity level at different points of time.Agriculture productivity measures that the area are performing less or higher compared with other regions nearby. The concept of agriculture productivity has been extremely used to explain the spatial organization and pattern of agriculture. Productivity is generally is considered into two directions, i) productivity of land ii) Productivity of infrastructure, productivity of land is linked with infrastructure closely. Productivity of land is an important factor of agriculture since it is permanent fixed factor among the three categories of input land, labor and capital. Agriculture productivity of land explained by production of crops in terms of yield per unit.Spatial analysis of agriculture productivity is very important because it can highlight the structure and pattern of production. (Dharmasiri, 2009) Agriculture productivity is defined by several researchers with their views and discipline. Agriculture productivity may be defined as the ‘ratio of index of agriculture output to index of total input used in farm production’ (shafi, 1984)In geography agriculture productivity is defined on “output per unit of input” or “output per unit of land area”. “Yield per unit” should be considered to indicate agriculture productivity (Singh and Dhilion, 2000).

Several studies have been made to measure the agriculture productivity and spatial pattern using different methods. Kendal used factor analytic approach and calculated latent roots to assign weight of individual crop production for the assessment of agricultural productivity patterns which emerged in England.Stamp (1958) suggested a method for measuring the agricultural productivity by converting the total agricultural productivity in calories. Shafi (1960) applied kendal ranking coefficient method for measuring the agriculture efficiency in the state of Uttar Pradesh. W. Tadesse and E Bekele (2001) attempted factor analysis to determine the relationship between the yield components and morphological characters of several genotypes of grasspea.

 

Review of Literature

In Agriculture study so many researches done in the past decade. In this section we are collecting some important research papers and present their findings and techniques.Jahan mohan et al., (2005) in his study found that area and production of cereal crops registered a negative trend in the entire agro climatic zone with the exception of western and southern zone that showed positive trend in the area of cotton. Dharmasiri (2009) had applied average productivity index to measure the agriculture productivity in Srilanka. Kalaivani et al. (2010) used Compound Growth Rate to measure the growth actions of area, production and yield for selected crops in Tamilnadu. In his study maize was recorded a positive trend in Tamilnadu. Sakthi mandal (2012) studied the spatial variation of agriculture productivity using Z score model and categorized the blocks as Very high, High, Medium and Low of agricultural productivity in south 24 PGS, Districts of West Bengal. Muthumurugan et al. (2012) suggested composite index analysis to study the agriculture development of Tamilnadu, he classified the districts based on the index value as highly developed, medium developed and low developed. Pajer Mansaram Pandit (2012), his study revealed that road, railway, agricultural labor, bank, good climate and modern technology are the important factors in the agriculture development of Nashik District. A S Rotaru et al. (2012) demonstrated factor analysis in agriculture by downsizing the data with various crops in the regions of Romania. Shafiquallah (2013) applied the technique Z-score model to study regional disparities and agricultural development in Uttar Pradesh.

 

Study Region

Tamilnadu has been classified into seven Agro Climatic Zones based on soil characteristics, rainfall distribution, irrigation pattern, cropping pattern and other social characteristics. The present study we are chosen North Western Agro climatic zone of Tamilnadu. This zone comprises the revenue districts of Dharmapuri, Krishnagiri, Salem and Namakkal. This zone is situated between 11and 12⁰55^’north latitude and 77⁰28^' and 78⁰50^’ east longitude; this zone covered an area of 16,150 sq km equivalent to 12.4 percent of the state area. Out of total geographical area of 17.31 lakh hactares, 8.01 lakh hectares are cultivated. The Average annual rainfall of the zone is 878 mm. This zone enjoys rainfall from both South West and North West monsoon seasons. The average temperature ranges maximum from 20° C to 42° C and minimum from 10°C to 31°C. Paddy, Maize, Ragi, Cumbu, Tapicaco and Sugarcane are the major crops cultivated in this zone. The main objective of this paper is to analyze the agriculture productivity, crops yield and spatial pattern and classification of North-western agro-climatic zone of Tamil Nadu.

 

Database and Methodology

The agriculture productivity data published by Department of Economics and Statistics, Chennai, Statistical Hand Book of Tamil Nadu was considered as database.The data consists of crops of each agro climatic zone for the time period of ten years (from 2003 to 2012).Among the listed crops, number of crops varied over the study period owing to removal of those crops for which the required data are not available. In this study, 15 crops are carefully chosen among the many that had been used in previous studies. These 15 mostly cultivated crops are chosen to asses agro productivity performance in the zone, they are Paddy, Cholam, Cumbu, Ragi, Maize, Bengal gram, Red gram, Green gram, Black gram, Horse gram, Tapioca, Ground nut, Gingely, Coconut and Sugarcane.

 

Data mining Techniques

Data Mining or Knowledge Discovery in Databases (KDD) is the process of discovering previously unknown and potentially useful information from the data in databases. In the present context data mining exhibits the patterns by applying few statistical data mining techniques namely, factor analysis, k-means clustering, Discriminant Analysis and Geographical Information System (GIS). As such KDD is an iterative process, which mainly consist of the following figure on the data collected;

 

 

Of these above iterative process 4 and 5 are most important. If appropriate techniques are applied in 5, it provides potentially useful information that explains the hidden structure.This structure discovers knowledge that is represented visually to the user, which is the final phase of data mining.

 

Factor Analysis

Factor Analysis is a generic term for a family of statistical techniques concerned with the reduction of a set of observed variables in terms of a small number of latent factors. The primary purpose of Factor analysis is data reduction, reducing data complexity by reducing the number of variables being studied. It has been developed for analyzing relationship among a number of measureable entities. It describes the covariance relationship among many variables in terms of few underlying by unobservable, random quantities called factors. Factor model is motivated by the following arguments.

1. To identify those variables with a particular group that are highly correlated among them but have small correlations with variables in different groups.
2. To reduce a large no of variables to a small number of factors.

 

Factor Extraction

There are several criteria for the number of factors to be extracted. Among the different extraction methods, the most widely used method; the principal component analysis is applied in this study to assess the performance of yield in the specified region.The eigen value greater than one rule suggested by Kaiser has been used in this study to extract the factors.

 

Factor Rotation

After deciding the number of factors extracted, the next step is to determine the method of Rotation. The rotation is used to simplify the factor structure and to achieve a more meaningful and interpretable solution. In the present study, factor analysis is initiated to uncover the patterns underlying crops variables.In factor extraction method the number of factors is decided based on the proportion of sample variance explained.Orthogonal rotations such as Varimax and Quartimax rotations are used to measure the similarity of a variable with a factor by its factor loading (Everitt and Dunn, 2001; Hair, Black, Babin andAnderson, 2010).

 

Non- Hierarchical Clustering Techniques

Cluster analysis is a multivariate method which aims to classify a sample objects on the basis of a set of measured variables into a number of different groups such that similar subjects are placed in the same group. Cluster analysis encompassed a number of different algorithms and methods for grouping objects of similar kinds into respective categories. The goal of clustering algorithm is to group the objects into a set of meaningful subclasses. This algorithm is used to discover structure in data without providing an explanation and determine the grouping of data themselves. In the present study one of the popular clustering algorithms suggested by Mac Queen (1967) known as k-means is used to identify the k-classes in the data set. The k-means method follows the following steps.

Step 1:Specify the number of clusters and, arbitrarily or deliberately, the members of each cluster.

Step 2: Calculate each cluster's centroid, and the distances between each observation and centroid. If an observation is nearer the centroid of a cluster other than the one to which it currently belongs, re-assign it to the nearer cluster.

Step 3: Repeat Step 2 until all observations are nearest the centroid of the cluster to which they belong.

Step 4: If the number of clusters cannot be specified with confidence in advance, repeat Steps 1 to 3 with a different number of clusters and evaluate the results

 

Multivariate Discriminant Analysis

Multivariate Discriminant function analysis builds a predictive model for group membership. The model is composed of a discriminant function based on linear combinations of predictor variables. Those predictor variables provide the best discrimination between groups. A discriminant score can be calculated based on the weighted combination of the independent variables. The main purpose of the discriminant analysis, to maximally separated the groups, to determine the most economical way to separate the groups and finally, to discard the variables which are little related to group distinctions. It is similar to regression analysis. This is achieved by the statistical decision rule which maximizes the between group variance relative to the within group variances. The discriminant ananlysis derives the linear combination from an equation that takes the following form

  , where, - Discriminant score, - discriminant weights and -Independent variables.In this research paper, we are using MDA for classifying the agriculture yield data in the North West Zone of Tamilnadu.

 

Geographical Information System

A Geographic Information System (GIS) is a system designed to capture, store, manipulate, analyze, manage, and present all types of geographical data. The first known use of the term "Geographic Information System" was by Roger Tomlinson in the year 1968 in his paper "A Geographic Information System for Regional Planning". GIS or spatial data mining is the application of data mining methods to analyze spatial data. Data mining, which is the partially automated search for hidden patterns in large databases, offers great potential benefits for applied GIS-based decision making. In the present study, GIS map is used to exhibit groups graphically and judge the nature of overall performance of the agriculture yield. A brief step-by-step algorithm to classify the crops during each of the study period based on their overall performances is described below:

Step 1: Factor analysis is initiated to find the structural pattern underlying the data set.

Step 2: k –means analysis is used to partition the data set into k-clusters using the original parameters in Step 1 as input.

Step 3: Discriminant analysis is then performed with the original variables by considering the groups formed by the k-means algorithm.

Step 4: GIS districts map is drawn with the standardized canonical discriminant function and centroid values.

 

Results and Discussions

Table 1: No. of years in each group

    1 – HYD, 2 – MYD, 3 – LYD.

 

Table 2: Eigen Values and Percentage ofVariance explained by factor

 

 

Table 3: Variable in Rotated Factors(Dharmapuri )

 

Table 4: Variable in Rotated Factors (Krishnagiri )

 

Table 5: Variable in Rotated Factors(Namakkali )

 

Table 6: Variable in Rotated Factors(Salem )

 

Table 7: Wilk’s Lambda

 

Table 8: Classification Results^*

      ^* 98.7 % original grouped classes correctly classified

 

Figure 1: Classification Map of Dharmapuri                                            Figure 2: Classification Map of Krishnagiri

 

 

                                  Figure 3: Classification Map of Namakkal                Figure 4: Classification Map of Salem

 

In Table 1 provides the groupings done by cluster analysis. Figure 1 to 4 shows the groupings of crops into 3 clusters for each districts of the study period in Table 1, we rated the members in the first cluster as High Yield District (HYD), and the second as Moderate Yield District (MYD) and the third as Low Yield District (LYD).Crops belonging to HYD category are the ones that perform better than those of MYD and LYD.Similarly the crops belonging to MYD category are superior to those of LYD, indicating the members in the category LYD are at a low report in terms of the crops yield considered in the present analysis. Table 7 shows that Wilks' lambda is a measure of how well each function separates cases into groups. Smaller values of Wilks’ lambda indicate greater discriminatory ability of the function. The associated chi-square statistic tests the hypothesis that the means of the functions listed are equal across groups. The small significance value indicates that the discriminant function does better than chance at separating the groups. In this study, Salem district is significantly different fromother districts. The classification matrix is reported in table 8. The pictorial representation of GIS mapping is drawn using the standardized discriminant functions evaluated at the group centroids. From the GIS maps in Figure 5, it is evident that the three groups of rated crops are very well separated and represented in the GIS maps for four districts. The Spatial pattern of districts (Figure 5) shows that Salem district get High Yield, Krishnagiri and Namakkal get Moderate Yield and Dharmapuri gets Low yield performance during the study period. The low yieldmay be duerong>