Lung Cancer with Prediction Using Dbscan

Abstract

Data mining in healthcare is an emerging field of high   importance   for   providing   prognosis   and   a   deeper understanding   of   medical   data.   Healthcare   data   mining attempts to solve real world health problems in the diagnosis and treatment of diseases. Researchers are using data mining techniques in the medical diagnosis of several diseases such as  diabetes,  stroke,  cancer  and  heart  disease.  Lung  cancer, also  known  as  lung  carcinoma,  is  a  malignant  lung  tumor characterized  by  uncontrolled  cell  growth  in  tissues  of  the lung. This growth can spread beyond the lung by the process of  metastasis  into  nearby  tissue  or  other  parts  of  the  body. Most  cancers  that  start  in  the  lung, known  as  primary  lung cancers, are carcinomas. The two  main  types are small-cell lung  carcinoma  (SCLC)  and  non-small-cell  lung  carcinoma (NSCLC).  symptoms  of  lung  cancer:  thready  pluse,  wiry pulse, soft pulse, thin fur, yellow fur, white fur, obesity,  red tongue,  deep  red  tongue,  pale  tongue,  internal  bleeding, abdominal distension, stomachache, constipation, loose stool, cough,   serious   cough,   occasional   cough,   phlegm,   little phlegm,  phlegm  that  is  hard  to  cough  up,  dry  mouth,  chest pain,  oppression  in  chest,  sweating,  fever,  weakness,  weak legs,  rash,  oral  ulcer.  As  there  is  a  big  growth  in  large volume   of   data   now   days,   this   will   create   a   need   for extracting  meaningful  data  from  the  information.  From  the various biomedical datasets, cancer is the widest disease that has  killed  human  life  over  7  million  every  year  and  lung cancer  among  them  is  nearly  17%  of  moralities.  Previous research  works  show  that  survival  rate  of  patients  affected with  cancer  is  larger  and  higher,  when  compared  to  the diagnosed  at  the  initial  stage,  Lung  cancer  is  the  most historic  data  and  dependent  disease  in  for  early  diagnosis. This has created the researcher to use data mining technique for early diagnosis of lung cancer in stage 1.There has been an increase in survival rate to about 70% at the early stage of detection, when tumor is not spread. Pre- existing techniques The  five  year  survival  rate  increases  to  70%  with  the  early detection  at  stage  1,  when  the  tumor  has  not  yet  spread. Existing     medical     techniques     like     X-Ray,     Computed Tomography (CT) scan, sputum  cytology analysis and  other imaging techniques not only require complex equipment and high  cost  but  is  also  proven  to  be  efficient  only  in  stage  4, when the tumor has metastasized to other parts of the body. Our   proposed   work   involves   the   uses   of   data   mining technique  used  in  classification  of  lung  cancer patients and the  categorization  of  stage  to  which  it  belong  positive.  The work is based on early diagnosis of prediction of lung cancer which   suggests   the   doctors   in   treating   the   patients   for increasing the survival rate of the human. For Early stage of prediction  there  are  so  many  data  mining  techniques  are there. I studied prediction using DBSCAN Algorithm