Logistic Regression with Elliptical Curve Cryptography to Establish Secure IoT

Nowadays,Wireless Sensor Network(WSN)is a modern technology with a wide range of applications and greatly attractive benefits,for example,self-governing,low expenditure on execution and data communication,long-term function,and unsupervised access to the network.The Internet of Things(IoT)is an attractive,exciting paradigm.By applying communication technologies in sensors and supervising features,WSNs have initiated communication between the IoT devices.Though IoT offers access to the highest amount of information collected through WSNs,it leads to privacy management problems.Hence,this paper provides a Logistic Regression machine learning with the Elliptical Curve Cryptography technique(LRECC)to establish a secure IoT structure for preventing,detecting,and mitigating threats.This approach uses the Elliptical Curve Cryptography(ECC)algorithm to generate and distribute security keys.ECC algorithm is a light weight key;thus,it minimizes the routing overhead.Furthermore,the Logistic Regression machine learning technique selects the transmitter based on intelligent results.The main application of this approach is smart cities.This approach provides continuing reliable routing paths with small overheads.In addition,route nodes cooperate with IoT,and it handles the resources proficiently and minimizes the 29.95%delay.

Sensor data compression based on MapReduce

A compression algorithm is proposed in this paper for reducing the size of sensor data. By using a dictionary-based lossless compression algorithm, sensor data can be compressed efficiently and interpreted without decompressing. The correlation between redundancy of sensor data and compression ratio is explored. Further, a parallel compression algorithm based on MapReduce [1] is proposed. Meanwhile, data partitioner which plays an important role in performance of MapReduce application is discussed along with performance evaluation criteria proposed in this paper. Experiments demonstrate that random sampler is suitable for highly redundant sensor data and the proposed compression algorithms can compress those highly redundant sensor data efficiently.