Probabilistic Threshold Query on uncertain data using SVM

Ms.Kavita Shevale, Prof.Gajanan Bhole


Data needed for fired query is available on the internet only concern is how effectively and efficiently it is delivered to the end user. This task is not easy because size of data is rapidly increasing and cost that we spend on correct data should be lesser than the value of data to be searched.
There is one more very important factor to be considered is uncertainty of data. Uncertainty of data is nothing but percentage of correctness in the result. Uncertainty of data may cause a hurdle in searching correct or desired data. To answer this type of query probabilistic approach is useful where the extent of accuracy is calculated. This accuracy calculation using probabilistic approach is very important to decide usefulness of the data. This may differentiate between very useful data a hardly useful data.
Paper concentrate on the probabilistic approach where the Support Vector Machine is utilized for the classification of a data; experimentally it has been proved that approach utilized delivers superior results than the approach where Enhanced Learning Machine is used.


Probability Threshold query, Support Vector Machine, Enhanced Learning Machine

Full Text:




  • There are currently no refbacks.

Copyright (c) 2017 International Journal of Advanced Research in Computer Science