International Journal of Civil Engineering and Technology (ijciet)

1. INTRODUCTION 
In recent years, as there exists an increasing attention towards research in the field of artificial 
neural networks which acts an attempt for modeling the capabilities in processing the data 
with essential properties. Even though there exists many models but it has become bottleneck 
for all other text editors to process the data in an efficient manner. Neural networks 
contemplate for an approach for problem of computation.
Main intention of this paper is to provide essential capabilities of analyzing and predicting 
the massive applications for handling the huge amount data efficiently. The processing of data 
is done to improve the accuracy the predicting data with adding the essential properties. As 
we turn up with experimenting the data along with practical approach uses strategies of 
learning and methods of data processing. Main objective is to serve two goals which includes 
as to accomplish neural network predicting results possible to the one who has been generated 
by super computer and secondary to increase the accuracy of predicting the huge amount of 
data. 

---

Analysis of Big Data with Neural Network
http://iaeme.com/Home/journal/IJCIET
212 
[e-mail] 
Big data analytics mainly involved in collecting huge amount of data from different 
sources and managing the data in such a way that depleted by analysts and finally is delivered 
to the useful organizational business. The process includes with converting huge assortment 
of data into structured and unstructured raw data which has been collected from different 
sources. The huge assortment of data is been collected for better business understanding and 
in order to process the data which is to be prepared for modeling the predications and then 
continued for evaluation.
In this the data is sampled nothing but to convert into data set for modeling. The data set 
is large volume of data which provides sufficient amount of information to retrieve in an 
efficient manner and it is explored for better understanding to overcome abnormalities with 
the help of data visualization. Then we focus on preparing the data for modeling in which to 
come up with a desired output as an evaluation is planned to come up with an accurate 
possibility of data.
The data processing module includes creation of data set design and the data set design 
leads towards two important portions of data which is 
•
Trained data set. 
•
Evaluation of data set. 
The trained data set consists of data which is to be trained within the huge organization of 
data which includes data in the form of rows. As each row consists of input vector id, input 
vector and targeted output. The input vector id is designed unique for each row in which 
includes input vector data in the form of dimensions which is followed up by targeted output 
in which results either true or false.
The Evaluation data set is the data which to be evaluated to improve the accuracy among 
different text editors. The data is been represented in the form of rows. As each row consists 
of id and input vector, where the id is an unique id to be used to identify the related input 
vector and input vector is the data which is represented in the form of 319 dimensions purpose 
is to generate predictions to evaluate the trained data set.
The input dataset is generated in the form of CSV(Comma Separated Values ) format 
which is to be loaded in the form of matrix. The data set has input vectors along with vector id 
and targeted output as discussed earlier. Here we come up with constructing two matrices 
from the trained data set which includes input data matrix and targeted data output matrix 
which is been included for huge millions of rows of each vector information. The trained data 
set is collected to predict the output which is done based on evaluation data set which ranges 
in the decimal format as 0 or 1 and separation is made between vector id and input vector to 
limit the origin which is used for accurate predictions which is to be saved using vector id 
later on after the evaluation process.