The paper addresses the security issues in private scalar product protocols, which are crucial for privacy-preserving data mining. It demonstrates that two previously proposed protocols, one by Vaidya and Clifton and another by Du and Atallah, are insecure. The attacks show that these protocols can be exploited to reveal private input vectors with high probability after a single execution. To address this, the authors propose a new private scalar product protocol based on homomorphic encryption, proving its security under standard cryptographic assumptions. The protocol is designed to be efficient, with optimized communication and computation for both parties. The paper also discusses practical considerations, including optimization techniques for binary vectors and batch computation, and compares the proposed protocol with other related work. The authors conclude by highlighting the importance of secure scalar product computation in privacy-preserving data mining and the practical efficiency of their proposed protocol.The paper addresses the security issues in private scalar product protocols, which are crucial for privacy-preserving data mining. It demonstrates that two previously proposed protocols, one by Vaidya and Clifton and another by Du and Atallah, are insecure. The attacks show that these protocols can be exploited to reveal private input vectors with high probability after a single execution. To address this, the authors propose a new private scalar product protocol based on homomorphic encryption, proving its security under standard cryptographic assumptions. The protocol is designed to be efficient, with optimized communication and computation for both parties. The paper also discusses practical considerations, including optimization techniques for binary vectors and batch computation, and compares the proposed protocol with other related work. The authors conclude by highlighting the importance of secure scalar product computation in privacy-preserving data mining and the practical efficiency of their proposed protocol.