Performance Analysis on Human Activity Detection using KNN and Random Forest
Sai Narendra L1, Samuel Kiran2, Naga Brahmani K3, Vamsidhar E4
1Sai Narendra L, Department of Computer Science Engineering, KLEF, Vaddeswaram (Andhra Pradesh), India.
2Samuel Kiran K, Department of Computer Science Engineering, KLEF, Vaddeswaram (Andhra Pradesh), India.
3Naga Brahmani K, Department of Computer Science Engineering, KLEF, Vaddeswaram (Andhra Pradesh), India.
4Vamsidhar E, Associate Professor, Department of Computer Science Engineering, KLEF, Vaddeswaram (Andhra Pradesh), India.
Manuscript received on 01 May 2019 | Revised Manuscript received on 15 May 2019 | Manuscript published on 30 May 2019 | PP: 2817-2821 | Volume-8 Issue-7, May 2019 | Retrieval Number: G5533058719/19©BEIESP
Open Access | Ethics and Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Human Activity Recognition is a promising area being able to profit the human culture by making assistive types of progress so as to help old, unendingly incapacitated and besides for individuals with phenomenal prerequisites. Exact improvement insistence is attempting since human movement is amazing and exceptionally different. Making study performed around there has uncovered information tunneling algorithm are utilized for solicitation of exercises. Hybrid mining frameworks, Naive Bayes with SVM and C4.5 with Neural Network are wound up being productive in portraying the accelerometers looking at information. These datasets are having wide arrangement of occasion with many proceeds with qualities. Working up a classifier that get-together such information is as of not long ago a troublesome errand. Sporadic woods is known for accomplishing high precision all together. It’s quality in social occasion broad datasets is promising. This paper proposes a sporadic timberland based course of action display for social classifying/predicting the strategy for activities. Preparing information is pre-managed to accomplish consistency. Points of reference from preparing dataset are pulled in sporadic for n tests, and n choice tree are made. Thus, an emotional choice backwoods is worked for depicting begins based accelerometers information respects. To predict unlabeled exercise information, total of n trees is performed. Primer takes a gander at are composed to inspect the movement confirmation limit of the model the outcomes are separated and transcendent regulated solicitation structures. It is seen that the proposed model beat the other depiction methodologies in relative examination. The sorted out social event show is restricted to perform action confirmation regarding weight lifting works out. Human Activity insistence is can be related with some reality, human-driven issues.
Keyword: KNN, Random Forests, Machine Learning.
Scope of the Article: High Performance Computing