The continuously increasing research in the field of science, engineering, and technology has generated textual and mathematical data in huge amounts. The research in retrieval and searching of textual data achieved the state-of-the-art results while searching and retrieval of mathematical information is in the early stage of research and requires significant improvement. Motivated from the concept of formula embedding and term-document matrix, in this paper we have introduced the variable size formula embedding approach where the formula is transformed into the variable size vector. In a vector, each bit represents their occurrence and corresponds to their position in BPIT. The proposed approach has been tested on the Math type data of Math Stack Exchange of ARQMath task and the proficiency of the same are represented in terms of nDCG, MAP and Precision at 10 measures. The obtained results have shown that the approach of variable size formula embedding requires significant improvement to retrieve the syntactically and semantically similar formula.