Smriti Sehgal | Amity University, Noida (original) (raw)
Uploads
Papers by Smriti Sehgal
2021 8th International Conference on Signal Processing and Integrated Networks (SPIN), 2021
As we know that in every city the demand for water is increasing day by day, it is becoming a ver... more As we know that in every city the demand for water is increasing day by day, it is becoming a very huge challenge to upgrade the filtration plants to meet the requirements. To do this for a particular city, the first task is to know the upcoming water requirement of that city. Some traditional approaches require a great deal to manual work, and it is very challenging to assess the performance of these traditional approaches. This problem was resolved by several researchers who proposed various algorithms using the Machine Learning algorithms like the Traversal grey model, Least Square Support Vector Machine and Based on Process Neural Network. To predict the consumption of water with maximum performance, in our work we have conducted a comparative study of a range of machine learning algorithms with some attribute selection techniques, k-folds cross-validation, and pre-processing techniques to boost the accuracy of these models. The performance of each algorithm will be evaluated on two parameters using the same dataset which will provide a comparison for the effectiveness of individual algorithms. These two parameters are Mean Absolute Error and Average Difference of Estimated and Real Value. At the end after comparing the results we get to pick the best algorithm to solve this problem. The algorithms used are Least Square Support Vector Machine, Lasso, XGBoost, Ridge regression, and a Hybrid model proposed which we have proposed in this paper. The obtained results proves that the hybrid model is the most suitable for the prediction of water consumption.
International Journal of Computer Applications, 2013
Remotely Sensed Images are composite images consisting of large number of spectral bands, from el... more Remotely Sensed Images are composite images consisting of large number of spectral bands, from electromagnetic spectrum. Analysis and Implementation of such images is much complex processing and takes lot of time. Therefore, dimension of these images must be reduced before any complex operation is performed. Selecting bands, which have higher capability to discriminate between classes, is a process of reducing number of bands with minimum loss of information [1]. In this paper, Canonical Analysis (CA) is used for band selection based on its discriminating power for classification of various classes. CA is based on Fisher's Linear Discriminant Analysis which maximizes the distance of pixels between classes and simultaneously minimizes the distance between pixels in the same class [5]. It computes eigenvalues and eigenvectors of each band for all the classes. Based on these values, loading factor matrix is computed and the band with highest discriminating power is given highest priority. Band with less priority are not selected leading to reduction of size of the image. Results show that spectral bands 1, 3, 5 are selected using Canonical Analysis whereas bands 4, 3, 2 are selected using Principal Component Analysis from the same LANDSAT image.
Remotely sensed images have detailed stored information spreaded over many spectral bands coving ... more Remotely sensed images have detailed stored information spreaded over many spectral bands coving full Electromagnetic spectrum. This information needs to be carefully extracted based on the type of segmentation one is doing or on the type of objects to be classified. In this paper, segmentation of high resolution image is done through bi-level and multi-level thresholding techniques. For bi-level, traditional OTSU method is used and Differential Evolution with OTSU technique as its objective function is used for multi-level thresholding. Segmented results with both the techniques are shown and it is clearly seen that differential evolution with OTSU method yield better results.
2021 8th International Conference on Signal Processing and Integrated Networks (SPIN), 2021
As we know that in every city the demand for water is increasing day by day, it is becoming a ver... more As we know that in every city the demand for water is increasing day by day, it is becoming a very huge challenge to upgrade the filtration plants to meet the requirements. To do this for a particular city, the first task is to know the upcoming water requirement of that city. Some traditional approaches require a great deal to manual work, and it is very challenging to assess the performance of these traditional approaches. This problem was resolved by several researchers who proposed various algorithms using the Machine Learning algorithms like the Traversal grey model, Least Square Support Vector Machine and Based on Process Neural Network. To predict the consumption of water with maximum performance, in our work we have conducted a comparative study of a range of machine learning algorithms with some attribute selection techniques, k-folds cross-validation, and pre-processing techniques to boost the accuracy of these models. The performance of each algorithm will be evaluated on two parameters using the same dataset which will provide a comparison for the effectiveness of individual algorithms. These two parameters are Mean Absolute Error and Average Difference of Estimated and Real Value. At the end after comparing the results we get to pick the best algorithm to solve this problem. The algorithms used are Least Square Support Vector Machine, Lasso, XGBoost, Ridge regression, and a Hybrid model proposed which we have proposed in this paper. The obtained results proves that the hybrid model is the most suitable for the prediction of water consumption.
International Journal of Computer Applications, 2013
Remotely Sensed Images are composite images consisting of large number of spectral bands, from el... more Remotely Sensed Images are composite images consisting of large number of spectral bands, from electromagnetic spectrum. Analysis and Implementation of such images is much complex processing and takes lot of time. Therefore, dimension of these images must be reduced before any complex operation is performed. Selecting bands, which have higher capability to discriminate between classes, is a process of reducing number of bands with minimum loss of information [1]. In this paper, Canonical Analysis (CA) is used for band selection based on its discriminating power for classification of various classes. CA is based on Fisher's Linear Discriminant Analysis which maximizes the distance of pixels between classes and simultaneously minimizes the distance between pixels in the same class [5]. It computes eigenvalues and eigenvectors of each band for all the classes. Based on these values, loading factor matrix is computed and the band with highest discriminating power is given highest priority. Band with less priority are not selected leading to reduction of size of the image. Results show that spectral bands 1, 3, 5 are selected using Canonical Analysis whereas bands 4, 3, 2 are selected using Principal Component Analysis from the same LANDSAT image.
Remotely sensed images have detailed stored information spreaded over many spectral bands coving ... more Remotely sensed images have detailed stored information spreaded over many spectral bands coving full Electromagnetic spectrum. This information needs to be carefully extracted based on the type of segmentation one is doing or on the type of objects to be classified. In this paper, segmentation of high resolution image is done through bi-level and multi-level thresholding techniques. For bi-level, traditional OTSU method is used and Differential Evolution with OTSU technique as its objective function is used for multi-level thresholding. Segmented results with both the techniques are shown and it is clearly seen that differential evolution with OTSU method yield better results.