There are number of cloud data owners who outsource their data to third party data mining servers in order to get frequent item sets from their datasets. But the main concern arises here is the security that the client of low processing power cannot substantiate correct mining result. So we are considering the server that is probably untrusted and attempts to elude from verification by using its prior knowledge of the outsourced data. So in order to verify these kinds of servers we propose efficient probabilistic and deterministic verification approaches that analyze whether the server sends correct and complete frequent item sets of our dataset. The first approach is probabilistic approach that can catch erroneous results with high probability and second deterministic approach measures the result correctness. Here we also design efficient verification methods for both situations that the data and the mining set are updated. In order to evaluate our results we are using various datasets and developing our project using JAVA technology.