A data mining project
$30-250 USD
着払い
It is based on the paper published by Nigam and Ghani in 2000. The dataset consists of some labelled data and large unlabeled data. The dataset is divided into two equal halfs A,B and Naive Bayes algorithm is used to develop a classifier model on each half. The classifier A is used to predict a class for the unlabeled data in A and these labels are also taken as the class labels for the other set unlabeled data i.e B. Now, the Naive Bayes is used on B and the whole labeled data is given as training. Then, this B classifier is again used to classify the unlabeled data which was not previously labelled. Now the predicted class labels are used in the place of previous labels done by A and these labels are also kept as A's labels. Now a classifier is again developed by new class labels in A and again predicting should be done. These new labels are also kept in B and again a model is developed in B. And the whole process continues until two converges.
プロジェクトID: #8937996
プロジェクトについて
アワード:
Our Software Company has 8+ years experience in Java; completed 270+ projects of GIS, Network, Database, Game, UML Diagram, etc. Recent Project : https://www.freelancer.com/projects/Java/Data-analysis-developmen もっと
12人のフリーランサーが、平均$161 で、この仕事に入札しています。
Hi, I have done serveral such data mining projects where I used classifiers such that Naive Bayes or WEKA for classifying of unlabeled data. I can do this task of predicting unlabeled data using WEKA. Thanks, Ra もっと
Hey i would the chance to talk some more about the project and what it is that you would need me to do. Contact me on chat so we can talk some more.
This is a very good example of data mining problem can be done with weka as a data mining tool.i do this problem easily.
Hi, I am an excellent statistician and R coder and can efficiently implement the Naive Bayes classifier for you. Feel free to ask any questions. I hope to work with you. John.