Velagapudi Ramakrishna Siddhartha Engineering College

About: Velagapudi Ramakrishna Siddhartha Engineering College is a based out in . It is known for research contribution in the topics: Computer science & Antenna (radio). The organization has 1307 authors who have published 1155 publications receiving 6163 citations.

Cancer Prediction Using Novel Ranking Algorithms and Machine Learning

Abstract: Cancer is the second leading cause of death globally. Especially, breast cancer is the most problematic cancer with more death rates. In this paper, we proposed a novel fusion classifier model based on a combination of various machine learning algorithms to improve accuracy. First, the base level models are trained and then we applied a ranking based algorithms for predicting final accuracy. The proposed model is tested on two different cancer datasets from UCI, Kaggle repository. The experimental results on two different datasets shown the effectiveness of the proposed framework. We used Python for implementing all our experiments.

Face Expression-Based Result Prediction in Talent Shows Using Deep Learning

Abstract: In this world, where people try to keep their likes or dislikes to one self. It is human nature to tend to grow curious to know the opinions of their performance if it is a talent show or their attitude regarding proposal or if it is a business meeting, etc. These are a few basic ideas where we tend to grow impatient to know the result or judgement, there are various other areas which we encounter in our daily life where knowing the outcome beforehand would do no harm. From that thought only evolved, in our project, we use deep neural networks to study one’s facial expressions to predict the possible outcome. Here, a new dataset of facial expressions of a judge from a talent show is collected. This dataset is used to train the convolutional neural network, which can then be used to predict the result of a judge.

Speech Separation Using Deep Learning with MATLAB

Abstract: The essential aspect in the field of speech processing is to resolve the Cocktail Party Problem and to realize speech-based human–machine interaction. For speaker source separation, Deep Learning is used as a potent means in order to isolate the speakers using the binary masking method for two speakers. The Mask is estimated with a set of training data and validation data and subsequently applied for test data. The latter method is based on separating three speaker sources utilizing the same binary masking method which involves various threshold techniques utilized to isolate speakers. This article presents how masking is applied to two speaker model and extended to three speaker model. The threshold techniques are implemented in the real environment to verify the efficiency.

California Bearing Ratio of a Cohesive Soil Reinforced with Waste Tyre Fibres

Abstract: This paper investigates the influence of waste tyre fibres on the geotechnical properties of clayey soil used as a subgrade material. Laboratory California bearing ratio (CBR) tests were conducted for the unsoaked condition as a measurement of pavement performance under the normal rainfall. Three types of tyre fibres as TFA, TFB and TFC used in the present study were mixed in a proportion of 0.25, 0.5, 0.75 and 1% by dry mass of clayey soil. In each test, the specimen was prepared in accordance with the compaction characteristics achieved from the modified compaction energy. The test results reported that the compaction characteristics of cohesive soil were getting increased with an increase in fibre content. In general, regardless of the fibre size, the CBR values get increased with an increase in fibre content. The experimental results showed that clayey soil achieves five times higher CBR value for the addition of 0.75% TFC. From the test results, it can be concluded that the use of waste tyre fibres considerably reduces the pavement thickness and hence reduces the construction cost.