Aim: An author’s aim to develop a project which is a random forest is a proposal algorithm to improve precision in contrast to the Adaboost algorithm already exist for the spam detection. In this research it contains 20 iterations which can be represented as group 1 and group 2. Hence 10 iterations of group 1 had taken and 10 iterations of group 2 had taken of this research with a pretest power of 80% and alpha value is 0.05. There are 5608 training datasets and 11,560 datasets are used in this research. The 95% confidence interval of the difference is 3.2164. Results: The prediction score has been obtained using a novel random forest algorithm of the first accuracy value given 99.74% and Adaboost algorithm given 97.93% so the rest of the values have been written to find the mean accuracy. The mean accuracy has been predicted that random forest algorithms (97.86%) were higher than the Adaboost algorithm (95.84%). The accuracy of mean and standard error difference is 2.02300 and 0.56805. Hence the significant value from equal variance assumed is p = 0.003 which is <0.05. Discussion: It has considerably better output compared to the Adaboost algorithm moving average output which has an accuracy of 97.93% for the social spam detection. Conclusion: The execution of this project has been successfully completed then it observes results and it tells that a novel random forest algorithm and its accuracy score improved than Adaboost algorithm has achieved for the spam identification.