| Summary: | Numerous works focus on the data privacy issue of the Internet of Things (IoT) when training a supervised Machine Learning (ML) classifier. Most of the existing solutions assume that the classifier’s training data can be obtained securely from different IoT data providers. The primary concern is data privacy when training a <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-Nearest Neighbour (<inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN) classifier with IoT data from various entities. This paper proposes secure <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN, which provides a privacy-preserving <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN training over IoT data. It employs Blockchain technology with a partial homomorphic cryptosystem (PHC) known as Paillier in order to protect all participants (i.e., IoT data analyst <i>C</i> and IoT data provider <i>P</i>) data privacy. When <i>C</i> analyzes the IoT data of <i>P</i>, both participants’ privacy issue arises and requires a trusted third party. To protect each candidate’s privacy and remove the dependency on a third-party, we assemble secure building blocks in secure <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN based on Blockchain technology. Firstly, a protected data-sharing platform is developed among various <i>P</i>, where encrypted IoT data is registered on a shared ledger. Secondly, the secure polynomial operation (SPO), secure biasing operations (SBO), and secure comparison (SC) are designed using the homomorphic property of Paillier. It shows that secure <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN does not need any trusted third-party at the time of interaction, and rigorous security analysis demonstrates that secure <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN protects sensitive data privacy for each <i>P</i> and <i>C</i>. The secure <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN achieved <inline-formula><math display="inline"><semantics><mrow><mn>97.84</mn><mo>%</mo></mrow></semantics></math></inline-formula>, <inline-formula><math display="inline"><semantics><mrow><mn>82.33</mn><mo>%</mo></mrow></semantics></math></inline-formula>, and <inline-formula><math display="inline"><semantics><mrow><mn>76.33</mn><mo>%</mo></mrow></semantics></math></inline-formula> precisions on BCWD, HDD, and DD datasets. The performance of secure <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN is precisely similar to the general <inline-formula><math display="inline"><semantics><mrow><mi>K</mi></mrow></semantics></math></inline-formula>-NN and outperforms all the previous state of art methods.
|
|---|