A COMPARATIVE STUDY ON EFFECTIVENESS OF DIFFERENT STATISTICAL TECHNIQUES – WITH SPECIAL REFERENCE TO AUTOMOBILE BUYING DECISIONS

: Different sort of statistical tools and packages are used widely used for different cases. These case results are elaborately analyzed by researchers. Each and every statistical technique has their own strong points as well as weakness. This paper focuses on a in-depth analysis on the effectiveness of the different statistical techniques like factor analysis, correlation, regression and some of the mining techniques. The problem scenario taken for the testing the result is automobile buying decisions in Kanchipuram.


INTRODUCTION
Earlier many research papers are presented based on statistical techniques and their effectiveness. In our earlier papers [1] [2], SPSS 19.0 and Weka tool was effectively deployed and results were presented. The same paper has been the motive and it is extended in a different scenario. These sort of studies will definitely be a guidance for the persons in supply chain management right from the producer end down to the consumer end. All are get benefited particularly the dealers in this industry the retail end show room operators and finally the consumer. So in that way this study is more oriented to the social environment.

LITERATURE REVIEW
Dr. Ronald Mani & Mr. Debasis Tripathy [3] conducted a study from 300 customers using random sampling technique to understand the influencing factors for purchasing two wheeler buying behavior in Uttar Pradesh. The study reveals that customers are satisfied with Honda and Bajaj company vehicles.
K. P Najeemudeen & N. Panchanatham [4] conducted a survey from 681 two wheeler consumers in Malappuram district of Kerla to understand the information sources for procuring their choice of two wheelers. A. Martin Jayaraj [5] conducted a study from 128 respondents in Coimbatore district using descriptive in nature to understand the awareness, influencing factors and reasons to choose particular dealer for their choice of buying two wheeler vehicles. S.Suraj Basha & Dr.B.C.Lakshmanna [6] identified that personal factors have significant impact in purchasing decision. Dr. Priyadarshini Padhi [7] conducted a study from 110 consumers from Ganjam to meet out the objectives of the study. The result reveals that family members play a vital role in purchase decision. Effective sales promotion leads to selecting particular brand. A. Anandalakshmy & Dr. K. Brindha [8], conducted a study from 150 respondents in Coimbatore city using convenient random sampling method and result reveals that majority of the women prefer scooty pep plus vehicle for the reason that smooth in running. Also colour and model leads to preference in choosing the vehicles.

OBJECTIVES OF THE STUDY
The main objective of the study is to analyze  Effectiveness of the strength and weakness of various statistical tools.  To present the snapshot about the automobile sector in Kanchipuram to the supply chain management group.

METHODOLOGY
In our earlier paper different sort of target questions were presented to different consumer groups and data were obtained. These data were sorted by mean analysis. And over this, SPSS 19 and Weka tool were tested. In this extended work, methodology were adopted to collect data and to normalize it. Over this normalized result, the effectiveness of various techniques like factor analysis, correlation, regression and mining techniques are applied and it was deeply analyzed.

Factor Analysis
The unique feature of this research work is that for the authentication for producing factor analysis or in another way dimension reduction has to be proved beyond doubt. For that we have chosen KMO & Barlett's test, from the table 1 & 3, the significant value is less than 0.05 which indicates that we can proceed for dimension reduction.  Initially we chosen 10 factors, on applying the varimax rotation technique it has been reduced to 5 crucial factors. From the component 1, three factors have been taken say word of mouth, finance/loan arrangement and loyalty for analysis. If we crucially look into these factors all can be club together as "Benevolent attitude of the dealer". This component induces strongly towards customer for their buying decision. In component 2, availability, insurance and infrastructure, this can be clustered as "Access to facilities". So the customer really needs to have good infrastructure and availability at any point of time. Under component 3, we have infrastructure and exchange offers, this can be clubbed together as "Value addition benefits". Any customer is much happy if they gets good amount of money from their old two wheeler and if its offered as good exchange price, their buying decision are much attracted. In component 4, we have service support and company image, this can be grouped as "After sales care", customers are delighted when he/she has a amicable after sales care for their two wheelers. This is having more influenced in their buying decision. Component 5, we have affordable price which can be put as "Motivational factor". This factor is named for the reason there is a say, "No other better motivation than money", money is the best motivation factor, so if customer gets a decent saving and quality product is afford to him/her of a competitive price he/she is much happy. Finally we conclude that we have 5 components, benevolent attitude of dealer, access to facilities, value addition benefits, after sales care and motivational factor, by this approach we have clustered different factors under 5 components. From the table 4 we have reduced 12 factors into 5 components to understand the reasons to choosing particular vehicle. Component 1 can be named as "Trouble freeness", component 2 can be named as "Additional benefit factor", component 3 can be called as "Comfortability", component 4 can be clustered as "Technical supremacy" and component 5 can be called as "Value for money". The strength of factor analysis technique is to reduce number of variables into components and to identify groups of inter-related variables. The weakness is difficult to pick proper rotation and interpretation is heuristic approach.

Correlation
From the table 5, clearly indicates that there is a significant correlation and it is diagonally symmetrical. This table clearly signifies the correlation connectivity among various factors. Advantage of correlation method is that it shows the strength of relationship between two variables either positive or negative. A disadvantage is that it does not reveal cause and effect.

Regression Analysis
From the regression analysis table 7, significant value is less than 0.05, hence model is valid and proceed for further analysis. It is observed from table 8 factor such as affordable price, word of mouth and availability is having strong impact in buying decision. This table is particularly applied for choosing a particular company for buying their two wheeler vehicle. This regression table is formed with specific approach towards choosing the company. It should be appreciated that here the approach is more specific than generic. Advantage of regression analysis is to examine relationship between dependent variable and several independent variables. Disadvantage is it cannot be used in qualitative phenomenon.   9 it is understood that model is valid and it can be proceed for further analysis. It is found from table 10, factor as brand image, speed, luggage space and seat comfort having strong impact in buying specific two wheeler decision.   Table 11 depicts different classification methods such as naive bayes and bayes net, it reveals that naive bayes method is better than bayes net through correctly classified instances. Table 12 shows different clustering methods such as simple k-means and hierarchical. Cluster 0 belongs to motor bike category and cluster 2 belongs to scooters category. Cluster 0 <--MOTOR BIKE Cluster 1 <--No class

CONCLUSION
In this research article, various statistical techniques like factor analysis, correlation, regression and some of the mining techniques like classification and clustering have been applied for specific issue of knowing buyers attitude for two wheeler selection in and around Kanchipuram town.
The results are already discussed in detail. The researcher's has to appreciate the fact that each technique is having efficacy in certain critical factors. So much more such statistical tools can be engaged over more specific socio economic problems which will be more much useful for both researcher as well as for business people and also for customers. This research has to be appreciate for getting defined results for the specific problem by three different statistical methodology and mining techniques.