The stock market prediction is a lucrative field of interest with promising profit and covered with landmines for the unprecedented. The markets are complex, non-linear and chaotic in nature which poses huge difficulties to predict the prices accurately. In this paper, a stock trading system utilizing feed-forward deep neural network (DNN) to forecast index price of Singapore stock market using the FTSE Straits Time Index (STI) in t days ahead is proposed and tested through market simulations on historical daily prices. There are 40 input nodes of DNN which are the past 10 days' opening, closing, minimum and maximum prices and consist of 3 hidden layers with 10 neurons per layer. The training algorithm used is stochastic gradient descent with back-propagation and is accelerated with multi-core processing. A trading system is proposed which utilizes the DNN forecasting results with defined entry and exit rules to enter a trade. DNN performance is evaluated using RMSE and MAPE. The overall trading system shows promising results with a profit factor of 18.67, 70.83% profitable trades and Sharpe ratio of 5.34 based on market simulation on test data.
The interference issue is most vibrant on low-powered networks like wireless sensor network (WSN). In some cases, the heavy interference on WSN from different technologies and devices result in life threatening situations. In this paper, a machine learning (ML) based lightweight interference mitigation scheme for WSN is proposed. The scheme detects and identifies heterogeneous interference like Wifi, bluetooth and microwave oven using a lightweight feature extraction method and ML lightweight decision tree. It also provides WSN an adaptive interference mitigation solution by helping to choose packet scheduling, Acknowledgement (ACK)-retransmission or channel switching as the best countermeasure. The scheme is simulated with test data to evaluate the accuracy performance and the memory consumption. Evaluation of the proposed scheme's memory profile shows a 14% memory saving compared to a fast fourier transform (FFT) based periodicity estimation technique and 3% less memory compared to logistic regression-based ML model, hence proving the scheme is lightweight. The validation test shows the scheme has a high accuracy at 95.24%. It shows a precision of 100% in detecting WiFi and microwave oven interference while a 90% precision in detecting bluetooth interference.
Recent studies on indoor positioning using Wi-Fi fingerprinting are motivated by the ubiquity of Wi-Fi networks and their promising positioning accuracy. Machine learning algorithms are commonly leveraged in indoor positioning works. The performance of machine learning based solutions are dependent on the availability, volume, quality, and diversity of related data. Several public datasets have been published in order to foster advancements in Wi-Fi based fingerprinting indoor positioning solutions. These datasets, however, lack dual-band Wi-Fi data within symmetric indoor environments. To fill this gap, this research work presents the UTMInDualSymFi dataset, as a source of dual-band Wi-Fi data, acquired within multiple residential buildings with symmetric deployment of access points. UTMInDualSymFi comprises the recorded dual-band raw data, training and test datasets, radio maps and supporting metadata. Additionally, a statistical radio map construction algorithm is presented. Benchmark performance was evaluated by implementing a machine-learning-based positioning algorithm on the dataset. In general, higher accuracy was observed, on the 5 GHz data scenarios. This systematically collected dataset enables the development and validation of future comprehensive solutions, inclusive of novel preprocessing, radio map construction, and positioning algorithms.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.