Pentaip | Global Financial Services | Company Information | Stock Data | Financial Report | Earnings | Analysis | Quarter & Yearly Report
Deep Learning (DL) is a subfield of Machine Learning (ML) and mainly used to recognize any of the use case that involves extremely complex pattern or scenario, especially in a huge and complicated classification of database, which is far beyond human ability to conduct any analysis or research in highly efficient and effective manner without any assistance of computing power. To perform Deep Learning (DL), we need to understand the concept Neural Networks at the first place. Neural Networks is also called Artificial Neural Networks (ANNs) that simulate the architecture of a human brain neural networks in computer and emulate the function (or abstract pattern) in the logical thinking and decision-making process as a human brain does.
Although “Forward Propagation” or “Forward Prop” process in the neural networks could give us the highest value in the advantages of timesaving and probabilistic accuracy in investment decision making process. However, the algorithmic model of neural networks needs to go through so called “Training” process from time to time, in order to ensure our algorithmic model still intact or being relevant to current market condition, as the accuracy of the prediction of neural networks is highly depending on its weights and biases on each neuron (node) within the layer. The degree of accuracy of the generated output in “Forward Propagation “process is based on the indication of “cost”, which is the difference between the “generated output value” and the “actual output value”, that the smaller the difference in the both value indicates the higher degree of the accuracy of the generated output to the actual output. To train our algorithmic model, the basic method is called “Back Propagation” or “Back Prop” that works backwards from the right (Output Layer) to the left (Input Layer), which is the reversed process to “Forward Prop” (from the left to the right).
TDALM is specifically designed for our users, to perform the complicated backtesting over their trading algorithms with variety of the most specific (or customized) criteria inputs (or variables) involved and to keep track on the backtesting result of their various complex trading algorithmic strategies running into their system, especially in any kind of the trading algorithmic strategy that combines a few different types of technical indicators (e.g. Moving Average, MACD, RSI and so on) and different timeframes involved sometimes (e.g. 3.3min, 33 min, 333min and so on), which creates more complexity to cope with those variances derived from the variables (or criteria inputs) involved in the backtesting. And so, we are here not only to provide our users a solution to perform complicated backtesting and also keep track on those backetested results of their various trading algorithms in their system, that enables them to work on any strategy adjustment on the fastest way as the institutional traders always do, to adapt any market change. (Coming Soon)
Language is not only the medium of communication among ours as a human, and the way of spreading information from the one to another. In no doubt, intelligence is the key to win in the financial market, and those are strong in collecting various types of data, converting it to the meaningful information and selecting the most relevant information that use it as intelligence, could highly possibly trade (buy/sell securities) in the most certain opportunity for riskless gain. However, there are tons of information generated from various sources each day, which could be in the form of news article, analysis or opinion report, social media comments (e.g. Donald Trump Twitter's posting), news update video and so on, that would affect the market sentiment from time to time. To define and filter the meaningful information from tons of database at extremely fast pace for keeping the "freshness", would generally involves lots of effort and manpower, which may not be cost-effective in doing so. Therefore, we are here to provide our users a customized solution that enables our users to keep track on the meaningful information without delay from the extraction of tons of the irrelevant information with our NLPCM system at the fastest pace, and also being able to extract those digital content such as news video (in audio mode) into human readable textual form for further analysis, which might refer to the Natural language processing (NLP). NLP is a subfield of Artificial Intelligence (AI), and basically focuses on the study in the combination of computer science and linguistics that empowers our computer to understand and communicate in human natural language (or in textual form), mostly in the unstructured way.
FDALM is specifically designed for our users, to perform the backtesting over those well-known investment philosophies from any gurus such as Benjamin Graham (Security Analysis, 1934), Warren Buffett (Long Term Value Investing), Peter Lynch (Peter Lynch Fair Value), Eugene Fam & Kenneth French (Fama-French 5-factor model, 2014), Joel Greenblatt (Magic Formula Investing, 2005), Philip Arthur Fisher (Common Stocks and Uncommon Profits, 1958) and so on, which give our users the actual picture on their basic belief and understanding to any kind of investing principle and any comment from famous Gurus, based on the fact and figure in the historical track record and the recent performance report from our backtesting system. The main function of FDALM is to remove any superstition (or unrealistic belief) on those old fashion investment principles that are no longer relevant to the current market condition. As an example, if our user believes that buying cheap stocks at extremely low P/E (below 3x) and P/B (below 0.5x) is the effective value investing method that's able to generate the high return in the stock market, and then, our user could perform the historical backtesting over the overall return on buying the cheap stocks (low P/E and low P/B) in different geographical markets, to see whether the basic belief (or assumption) is still relevant to the actual market performance, in consideration of any further application of the fundamental belief in investing. (Coming Soon)
Only for limited invtation to our collaborative partner upon a special request in customized system development. (Coming Soon)