Recognizing human emotions by machines is a complex task. Deep learning models attempt to automate this process by rendering machines to exhibit learning capabilities. However, identifying human emotions from speech with good performance is still challenging. With the advent of deep learning algorithms, this problem has been addressed recently. However, most research work in the past focused on feature extraction as only one method for training. In this research, we have explored two different methods of extracting features to address effective speech emotion recognition. Initially, two-way feature extraction is proposed by utilizing super convergence to extract two sets of potential features from the speech data. For the first set of features, principal component analysis (PCA) is applied to obtain the first feature set. Thereafter, a deep neural network (DNN) with dense and dropout layers is implemented. In the second approach, mel-spectrogram images are extracted from audio files, and the 2D images are given as input to the pre-trained VGG-16 model. Extensive experiments and an in-depth comparative analysis over both the feature extraction methods with multiple algorithms and over two datasets are performed in this work. The RAVDESS dataset provided significantly better accuracy than using numeric features on a DNN.

With the advent of big data era, enormous volumes of data are generated every second. Varied data processing algorithms and architectures have been proposed in the past to achieve better execution of data mining algorithms. One such algorithm is extracting most frequently occurring patterns from the transactional database. Dependency of transactions on time and location further makes frequent itemset mining task more complex. The present work targets to identify and extract the frequent patterns from such time and location-aware transactional data. Primarily, the spatio-temporal dependency of air quality data is leveraged to find out frequently co-occurring pollutants over several locations of Delhi, the capital city of India. Varied approaches have been proposed in the past to extract frequent patterns efficiently, but this work suggests a generalized approach that can be applied to any numeric spatio-temporal transactional data, including air quality data. Furthermore, a comprehensive description of the algorithm along with a sample running example on air quality dataset is shown in this work. A detailed experimental evaluation is carried out on the synthetically generated datasets, benchmark datasets, and real world datasets. Furthermore, a comparison with spatio-temporal apriori as well as the other state-of-the-art non-apriori-based algorithms is shown. Results suggest that the proposed algorithm outperformed the existing approaches in terms of execution time of algorithm and memory resources.

Owing to accurate future air quality estimates, need for detecting the anomalously high increase in concentration of pollutants cannot be adjourned. Plentiful approaches were proposed in the past to substantially determine the abnormal conditions, but most of the statistical approaches were computationally expensive as well as ignored the false alarm ratios. So, a hybrid of proximity and clustering based anomaly detection approaches to identify anomalies in the air quality data is suggested in this work. Gaussian distribution property of real world dataset is utilized further to segregate out anomalies. The results depicted twofold advantages of our approach by efficient extraction of anomalies as well as increased accuracy by reducing the number of false alarms. Specifically, the presence of NO₂ concentration in air is investigated in this work considering its constant increase over decades as well as its inevitable health risks. Furthermore, spatio-temporal segments with anomalously high NO₂ concentrations for 14 residential, industrial and commercial areas of 5 cites of India are extracted out. To validate the results, a comparative analysis with existing approaches of anomaly detection as well as with two benchmark datasets is performed. Results showed that our method outperformed the existing methods of anomaly detection, when evaluated over metrics such as sensitivity, miss rate and false alarms. Further a detailed analysis of extracted anomalies as well as a detailed discussion about the factors responsible for such anomalies is presented in this work. This study is helpful in educating Government and people about spatio-temporal, geographical and economic conditions responsible for anomalously high NO₂ concentrations in air.

The advancements in deep learning have brought about crucial transformations in computer vision over the past two decades. Deep convolutional networks have found many applications in building fine-tuned models for implementation in vision-related tasks. Knowledge learned by deep learning models over enormous generic datasets can be transferred to be employed for much more specific tasks. In this work, are implementing the approach to provide health benefits to people. In the present work, we develop an application which help them in performing exercises without the help of a trainer and get instant feedback about the postures. We aim to make fitness accessible to all by removing barriers such as external hardware requirements and cost-based subscriptions. In this paper, we dive deep into the technical details about the application and the exact methodologies applied for building the same. Furthermore, results are evaluated after running the application over multiple scenarios and a comparative analysis is performed.