In heterogeneous datasets while used matching instances state-of-the-art instance matching approaches do not perform well. From the central part operation on direct matching these disadvantages should be generated. The direct matching involves a direct connection between instances at the source dataset and instances at the target dataset. If the overlap between the datasets is minimum direct matching is not applicable for that type of application. The big aim of this survey is resolving this problem by proposing a new paradigm called semantic similarity matching. The class of interest is defined as a class of instances from the source dataset. The class-based matching is defined as a set of candidate matches retrieved from the target. Measuring the semantic similarity between words is an important component in various web applications they are relation extraction, community mining, document clustering, and automatic metadata extraction. The usefulness of semantic similarity measures in these applications, accurately measuring semantic similarity between two words or entities. This survey proposes an empirical method to estimate semantic similarity using page counts and text snippets. Both should be retrieved from a web search engine for two words. The various word co-occurrence measures using page counts should be integrated with lexical patterns those should be extracted from text snippets. To identify the numerous semantic relations between two words a novel pattern extraction algorithm and a pattern clustering algorithm should be introduced. The proposed method outperforms various baselines compared to previously proposed SERMI approach on three benchmark data sets showing a high correlation with human ratings. Moreover, the proposed method significantly improves the accuracy in a community mining task.