Search

Stock Trading has been one of the most important parts of the financial world for decades. People investing in the share market analyze the financial history of a corporation, the news related to it and study huge amounts of data so as to predict its stock price trend. The right investment i.e. buying and selling a company stock at the right time leads to monetary benefits and can make one a millionaire overnight. The stock market is an extremely fluctuating platform wherein data is produced in humongous quantities and is influenced by numerous disparate factors such as socio-political issues, financial activities like splits and dividends, news as well as rumors. This work proposes a novel system “IntelliFin” to predict the share market trend. The system uses the various stock market technical indicators along with the company's historical market data trends to predict the share prices. The system employs the sentiment determination of a company's financial and socio-political news for a more accurate prediction. This system is implemented using two models. The first is a hybrid LSTM model optimized by an ADAM optimizer. The other is a hybrid ML model which integrates a Support Vector Regressor, K-Nearest Neighbor classifier, an RF classifier and a Linear Regressor using a Majority Voting algorithm. Both models employ a sentiment analyzer to account for the news impacting the stock prices which is powered by NLP. The models are trained continuously using Reinforcement Learning implemented by the Q-Learning Algorithm to increase the consistency and accuracy. The project aims to support the inexperienced investors, who don't have enough experience in investing in the stock market and help them maximize their profit and minimize or eliminate the losses. The developed system will also serve as a tool for professional investors to help and aid their decision making.

Selecting appropriate surrogate models is crucial. This article upgrades our learning-based surrogate selection paradigm (LEAPS2) to LEAPS2v2. Its key features include: modeling noisy and non-noisy data, avoiding noise-fitting surrogates, powerful data attributes, novel composite metric (Surrogate Quality Score) to assess surrogate accuracy and complexity, and many surrogates and datasets. For any dataset, LEAPS2v2 recommends 3 out of 36 surrogates. A LEAPS2v2 recommendation is successful if it recommends at least one of the top three surrogates based on Surrogate Quality Score. LEAPS2v2 was successful on more than 94 % non-noisy and 89 % noisy datasets. Moreover, it made successful recommendations on 14/16 real industrial datasets. Strong correlation was observed between a surrogate being a true best and being recommended. Our numerical analysis revealed that the best surrogates for noisy data are different from those for non-noisy datasets. No single surrogate outperforms others for all datasets, highlighting the utility of LEAPS2v2. Publication Status: Submitted to Computers & Chemical Engineering.

People detection and counting is considered as one of the important application in video surveillance. Various computer vision and deep learning based methods have been developed which aim to provide efficient and accurate people detection/counting results using frontal view data sets. Furthermore, there are many challenges which occurs while detecting people including occlusion, perspective distortion, variations in human body pose, size and orientation, these challenges effect the results of developed detection and counting models. In this work, a deep neural network approach i.e. SSD (Single Shot multi-box Detector) is explored for people detection and counting. SSD model is used for detection and counting of people from significantly different viewpoint i.e. top view. To the extent of our knowledge, this is the first attempt to use deep neural network based model for top view people detection and counting. Furthermore, the impact of frontal view trained SSD model on top view test images is also discussed. The experimental results show the effectiveness of deep learning model by achieving promising results with average TPR of 95% and TPR 94.42% for indoor and outdoor environments respectively.

Nowadays skin diseases are considered as one of the common cause of human illness, and if not detected or diagnosed at early stage, it can be more harmful. With advancement in computer vision and image processing various techniques have been developed for diagnoses of skin diseases. This work mainly focused on simple yet effective technique for segmentation of skin lesions. It is also robust for extracting the region of interest whose background and foreground color is bit similar. It automatically segments skin lesion of arbitrary shapes and sizes. The proposed technique is mainly based on three major modules namely preprocessing, segmentation and post-processing. In pre-processing different color models are used for extraction of single channel followed by filtering to remove noise. In the second module, segmentation is preformed using Harris corner detector, seed selection and region growing segmentation. Finally morphological operation is performed as post-processing comprising of region filling and dilation. The experimentation results shows the performance of proposed model by achieving accuracy of 95%.

In this paper we propose a probabilistic approach to synthesize an agent-based heterogeneous semantic model depicting population interaction and analyzing the spatio-temporal dynamics of an airborne epidemic, such as influenza, in a metropolitan area. The methodology is generic in nature and can generate a baseline population for cities for which detailed population summary tables are not available. The joint probabilities of population demographics are estimated using the International Public Use Microsimulation Data (IPUMS) sample set. Agents are assigned various activities based on several characteristics. The agent-based model for the city of Lahore, Pakistan is synthesized and a rule based disease spread model of influenza is simulated. The simulation results are visualized to produce semantic analysis for the spatio-temporal dynamics of the epidemic. The results show that the proposed model can be used by officials and medical experts to simulate an outbreak.