Classification of Low Birth Weight Baby Under Anthropometry uses Algorithms K-Means Clustering on Maternity Hospital

LBW infants with birth weight less than 2500 grams regardless gestation period. Low birth weight is the weight of a baby who weighed within 1 hour after birth. World Health Organization (WHO) since 1961 states that all newborns are underweight or equal to 2,500 g called low birth weight infant (low birth weight). According to WHO. Statistically, morbidity and mortality in neonates in developing countries is high, with the main causes is associated with LBW. To facilitate medical personnel in determining the risk of LBW. From the testing that has been done by the author, the k-means clustering algorithm has accuracy in classifying LBW babies by spacing the proximity between variables and the similarities in the test data,

geometry, mass, strength and characteristics of the human body in the form of shapes and sizes. Humans will basically have the shape, height and weight are different from one another, for the research was conducted to classify LBW infants by anthropometric using k-means algorithm. K-Means Clustering is a method of analyzing data or methods that perform data mining modeling process without supervision (unsupervised) and is one method of grouping the data to the system partition. The conclusions analysis who obtained the best algorithm to perform clustering in the above studies is the K-Means algorithm, for grouping clusters in K-means algorithm is better than the EM algorithm [4]. The calculation results show K-Means method produces better results than using DBSCAN. Based on these studies, the authors chose to use an algorithm K-Means clustering method in this study to classify low birth weight infants. So it can be found certain patterns that yield important information that support the process of checking LBW [1]. On this topic based on the above background can be taken the problem in this research is, how the results grouping LBW infants by anthropometric measurements using data mining with k-means clustering algorithm. The aim of this research is the grouping infants included in the diagnosis of LBW to help midwives and medicine in determining LBW infants groups based on the calculation of k-means clustering algorithm.

Data Mining
Data mining is a series of processes for adding additional value of a set of data in the form of knowledge that had been unknown to them manually [3].

K-Means Algorithms
K-Means clustering is a method of group analysis that led to the object of observation of partitioning N into K groups (clusters) where each object observation is owned by a group with a mean (average) nearby [5]. the steps do clustering with K-Means method as follows : a. Select the number of clusters k. b. Initialize k clsuster center with random manner. c. Data to calculate the distance to each centroid using distance formula Euclidean Distance. Here is the equation of Euclidean Distance: Where: D (i, j) = Distance Data to i to the cluster center j Xij = Data to i the data attributes to k Xkj = Data to j the data attributes to k d. Recalculate center of the cluster with the current cluster membership. Determine each object again wear the new cluster center. If the cluster centers do not change again the clustering process is completed. Or, go back to step c until the cluster centers do not change anymore.

Anthropomerty
Anthropometry is derived from the "anthro" means human and "metron" meaning measure. Anthropometric definitively declared as a study concerning the measurement of human body dimensions and design applications involving physical geometry, mass, strength and characteristics of the human body in the form of shapes and sizes. Humans will basically have the shape, height and weight are different from one another [9].

Low Birth Weight (LBW)
Birth weight is the weight of neonates at birth weighed within one hour after the body lahir.berat anthropometric measure the most important and most frequently used in newborns (neonates). Weight loss is used to diagnose normal or low birth weight infants [2].

WEKA Tools
Weka is an open source data mining applications based on java. This application was first developed by the University in New Zealand named the University of Waikato in 1994 [7].

Method
In this research using primary data, ie data Antropometeri birth of Maternity Hospital Indonesian. Here is the flow of research using k-means clustering algorithm. This step for fulfillment : From the picture above, the lines of inquiry that are used as follows: 3.1. Identification Problem The process of problem identification is performed to determine the problems and methods that would allow them to ditentuka points to classify infants into the category of LBW.

Data Collecting
In this research uses data Anthropometric birth of Maternity Hospital on Purwokerto Indonesian teritorial.

Data Preprocessing
At this stage, a data transformation process that is changing the nominal data into numerical data.

Using Clustering Method
To the use of clustering algorithms grouping Kmeans to get results based on data Anthropometric LBW infants.

Evaluated
In this phase all evaluated from the beginning to get the grouping which is processed using k-means clustering algorithm.

Result and Conclusion
The next stage is to conclude the results obtained from research based algorithm K-Means clustering that delivers results instagram social media use and value of students.

Results and Discussion
The sample used in this research were 1210 babies of 2016. Table 1 shows the data babies sample data.

Handling Missing Values
At the stage of preprocessing the data, the next step needs to be done after the data set is to perform the data processing to handle the data subject to a missing values. In the data LBW data are missing values experienced in column 640 are attribute to weight attributes and attribute column to which the attribute LK 985 (Round Head). Below is a table of 2 missing data values.

The use of clustering methods
After going through the preprocessing phase completed, the dataset start calculation and processed using weka apilkasi. The use of the above method is just to do grouping, the grouping obtained from the calculation iterations that will be performed by researchers. The data used in this study as many as 1210 data. At this stage, the author uses primary data to pengolaha data, using anthropometric data childbirth.

Conclusion
In this study, we concluded grouping of testing k-means clustering algorithm. From the testing that has been done by the author, then the algorithm k-means clustering to have accuracy in classifying LBW infants derdasarkan distance of proximity between variables and the similarities in the test data, such as results already in the know, that the baby is diagnosed LBW everything into the cluster group second, that there are 107 data or 9% of LBW dataset in which there are indications of LBW babies, while another group is a group of babies are normal and do not get into the group of LBW.