OPUS Collection:
http://hdl.handle.net/10453/19979
2016-12-13T02:10:42ZIn-depth optimization of stock market data mining technologies
http://hdl.handle.net/10453/66014
Title: In-depth optimization of stock market data mining technologies
Authors: Lin, Li
Abstract: Stock trading is a number of stocks to be exchanged from one trader to another
trader. It consists of a trader selling a number of stocks at a price and a volume,
and another trader buying the same stocks at the same price and the same volume.
Most traders want to buy a stock at a low price and to sell the stock at a high
price in order to make a profit. However, it is difficult to know whether the
current trading (buy/sell) price is low or high. Some researchers have presented
technical trading rules which are mathematical formulas with many parameters to
solve this problem, such as moving average rules, filter rules, support and
resistance, channel break-out rules, and so on. All these rules are based on
historical data to generate the best parameters and use the same parameters in
future trading to make a profit. When the parameters of a trading rule are set
properly, the trading rule can help the traders to make a profit (buy/sell at a
low/high price). Experiments have shown that technical trading rules are
profitable.
However, there are still some disadvantages and limitations to the technical
trading mies in real stock market trading. First, the technical trading rules do not
integrate domain knowledge (expert experiences and domain constraints, etc).
For example, some trading rules pattern maybe only generate three signals during
one year trading to get the most profit. However, the pattern is unreasonable and
it is unprofitable in future trading, because the pattern is only a mathematical
maximum, but it is impracticable in stock trading. Second, the output of a
parameter for a trading rule is only one single value. Sometimes, it may be a
noise so the trading rule is inapplicable in future trading. Third, present
algorithms to calculate parameters of trading rules are inefficient. Most trades are
performed through internet such that they can buy and sell stocks in online and a
trade is completed in a second. Real markets are dynamic such that trading rules
have to be updated all the time depending on changing situations (new data come
in, new parameters will be recomputed). Current enumerate algorithms waste too
much time to get new parameters. However, a one-second short delay in real
stock trading will lose the best trading chance. Fourth, when we evaluate the
performance of a stock, we need not only to consider its performance (profit and
return), but also to compare it to other stocks performance. At present, trading
rules do not compare to the other stocks performance when they are selected to
generate a signal, so the selected stocks or rules may be not the best ones. Fifth,
in stock markets, there are many stocks and many trading rules. The problem is
how to match and rank stocks and rules to combine a profitable and applicable
pair. However, trading rules do not solve this problem. Lastly, trading rule
techniques do not consider the sizes of investments. However, in real market
trading, different investments will result in a different performance of a pair.
We propose in-depth data mining methodologies based on technical trading rules
to overcome these disadvantages and limitations mentioned above. In this thesis,
we present the solutions to combat the existing six problems.
To address the first problem, we designed a domain knowledge database to store
domain knowledge (expert experience and domain constraints). During the
computing procedure, we integrated domain knowledge and constraints. We
observed the output more reasonable as we considered domain knowledge.
To address the second problem, we optimized a sub-domain output instead of a
single value, in the sub-domain all combinations of parameter can get a near-best
result. Moreover, in the sub-domain, some experienced traders can also set or
micro-tune parameters by themselves and a better performance is guaranteed.
To address the third problem, we adopted genetic algorithms and robust genetic
algorithms to improve the efficiency. Genetic algorithms and robust genetic
algorithms can get a near-optimal result in an endurable execution time, and the
result is also near to the best one.
To address the fourth problem, we applied fuzzy sets and multiple fitness
functions to evaluate stocks. Because many factors influence the performance of
a stock, it is necessary to create a multiple fitness function for genetic algorithms
and robust genetic algorithms.
To address the fifth problem, we built a stock-rule performance table to rank
stock-rule pairs and find the best matching pairs. The stock-rule pair results
showed that the ranked performance is better than that of randomly matched
pairs.
Finally, to address the sixth problem, we drew a graph of the relationship
between investments and number of stock-rule pairs to search maximal points,
and to decide the number of pairs for different sizes of investments.
In summary, the purpose of this thesis is to identify optimal methodologies in
stock market trading, to make more profit with less risk for investors. The
experimental results showed that the methodologies are more profitable and
predictable.
Description: University of Technology, Sydney. Faculty of Information Technology.2006-01-01T00:00:00ZMatlab/Simulink modules for modeling and simulation of power electronic converters and electric drives
http://hdl.handle.net/10453/65847
Title: Matlab/Simulink modules for modeling and simulation of power electronic converters and electric drives
Authors: Iyer, Narayanaswamy P.R
Abstract: Modelling and simulation of power electronic converters and electric drives play a vital
role in the academic curriculum and also in the industry. A number of modelling and
simulation tools are used to study the performance of power electronic converters and
electric drives. For the analysis and simulation of three phase electric drives, the three
(abc) axis to two (dq) axis transformation is used [I]. The results in the dq axis is
transformed to abc axis by suitable inverse transformation.
Over the past, several analog, hybrid and digital computers were used for simulation of
converter fed electric drives [2, 13, 29, 30, 33, 34, 35]. In the recent years, a number of
software packages have been developed to study the performance of power electronic
converters and converter fed electric drives [3 - 9].
SIMULINK developed by Mathworks Inc., USA. is one of the softwares used for
power electronic converters and electric drive simulation [3, 4, 11]. This software is
used for modelling the power electronic converters and electric drives discussed in this
thesis.
This thesis describes the interactive modelling of power electronic converters such as ac
to de, de to ac, de to de and ac to ac and ac drives such as the three phase IM and Six
Step Inverter fed PMSM, using the software SIMULINK. Unless specified otherwise,
the term "model" in this thesis refers to SIMULINK model. Interactive Library
Building Blocks are developed using SIMULINK for the above power electronic
converters. These library models are then used to develop PWM converters. The
models for well known PWM techniques such as Sine, HI, THI are presented. The
interactive model for a totally new PWM technique known as Clipped Sinusoid PWM
(CSPWM) is presented in this thesis. Where possible the results are compared with
literature references, by theoretically derived formula and also by Electronic Circuit
Simulation software..
Interactive Circuit Model of a Digital Gate Drive for a Three Phase 180 Degree mode
two level inverter using four line to one line multiplexer is presented and the results
compared with well known literatures on power electronics and also by experimental
verification.
Interactive system Models for three phase ac Line fed IM drive in all reference frames
using dqO voltage - current and flux linkage equations in state space are presented and
simulation results compared with the literature references. This is followed by various
system models for three phase Pulse Width Modulated Inverter fed IM drive.
Interactive system models for Six Step Continuous and Discontinuous current mode
inverter fed PMSM drives are presented and the results are compared experimentally,
by theoretically derived formula and also with the literature references.
Interactive system models for Buck Converter Switched Mode Power Supply (SMPS)
are given and the results compared with the literature references and also by electronic
circuit simulation.
Interactive system models for Three phase DCTLI and FCTLI are presented and the
result compared with literature references and also by theoretical derivations.
Harmonic analysis of six step continuous current mode two level inverter and three
phase three level inverter are presented in APPENDIX A. Experimental data and
MATLAB programs to calculate the parameters of the six step Lybotec inverter fed
PMSM drive in the laboratories of CEMPE are presented in APPENDIX B. The block
diagram schematic of the six step Lybotec Inverter in the laboratories of CEMPE is
provided in APPENDIX C. Some data sheets for selected integrated circuits are
provided in APPENDIX D. Comparison of the model performance of Power
Electronic Converters and Electric Drives presented in this thesis made with the
Electronic Circuit Simulation Software PSIM, MICROCAP8 and the SimPowerSystems
Block set of SIMULINK is presented in APPENDIX E. The list of publications from
this thesis is given in APPENDIX F.
Description: University of Technology, Sydney. Faculty of Engineering.2006-01-01T00:00:00ZDiversification philosophy and boosting technique for trade execution strategy
http://hdl.handle.net/10453/64809
Title: Diversification philosophy and boosting technique for trade execution strategy
Authors: Wang, Jiaqi
Abstract: This thesis explores the rationale and effectiveness of diversification across time
and strategies, which is an important philosophy for risk management in practice, in
the framework of developing trade execution strategies. In this thesis, the strategies
are defined as making a series of decisions based on real-time state variables over a
fixed period to achieve high reward and low risk with given resources. Trade
execution strategies are to make a series of decisions on how to place an order in
markets based on real-time market information over a fixed period to fill the order
with low cost and risk in the end.
In the 1st part, this thesis explores diversification across time. The research of
trade execution has shown that although limit order strategy achieves lower cost
than market order strategy does, it may incur nonexecution risk and miss trading
opportunities. This thesis proposes a strategy that reflects the idea of diversification
across time to improve the limit order strategy. In the 2nd part, this thesis explores
diversification across strategies. Techniques for implementing this idea have been
proposed to acquire strategies from a candidate strategy set and determine their
weights. For those techniques, the candidate strategy set normally only contains
finite strategies and the risk that they reduce is only measured by one specific
standard. This thesis proposes a technique that overcomes those drawbacks. In the
3rd part, the proposed technique is applied to improve trade execution strategies.
The strategy proposed in the 1st part is called DF (dynamic focus) strategy, which
incorporates a series of small market orders with different volume into the limit
order strategy and dynamically adjusts each market order volume based on two real-time
state variables: inventory and order book imbalance. The sigmoid function is
adopted to map the variables to the market order volume. Experiments show that the
DF strategy achieves lower cost and risk than the limit order strategy does.
The technique proposed in the 2nd part extends the key idea of the AdaBoost
(adaptive boosting) technique, which is discussed mostly in the supervised learning
field. It is named DAB (diversification based on AdaBoost) in this thesis. The DAB
technique adaptively updates the probability distribution on training examples in the
learning process, acquires strategies from a candidate strategy set and determines
their weights. Resources (e.g. money or an order) are allocated to each acquired
strategy in proportion with its weight and all acquired strategies are then executed
in parallel with their allocated resources. The DAB technique allows the candidate
strategy set to contain infinite strategies. Analysis shows that as the learning steps
increase, the DAB technique lowers the candidate strategy set's risk, which can be
measured by different standards, and limits the decrease in its reward.
The DAB technique is applied in the 3rd part to acquire DF strategies from a
candidate DF strategy set and determine their weights. The entire order is allocated
to each acquired DF strategy in proportion with its weight and all acquired DF
strategies are then executed in parallel to fill their allocated order. In this thesis, this
parallel execution is called BONUS (boosted dynamic focus) strategy. Experiments
support theoretical analysis and show that the BONUS strategy achieves lower risk
and cost than the optimal DF strategy and two simple diversification techniques do.
This thesis is contributed to both finance and computer science fields from the
theoretical and empirical perspectives. First, the proposed DF strategy verifies the
effectiveness of diversification across time through improving the existing trade
execution strategies. Second, the proposed DAB technique provides a flexible way
for implementing diversification across strategies to complement the existing
diversification techniques and enrich the research of the AdaBoost technique. Third,
the proposed DAB technique and BONUS strategy provide a flexible way to
improve trade execution strategies.
Description: University of Technology, Sydney. Faculty of Information Technology.2006-01-01T00:00:00ZDevelopment and evaluation of a CAI course in Information Technology for Life at Nakhon Pathom Rajabhat University, Thailand
http://hdl.handle.net/10453/64786
Title: Development and evaluation of a CAI course in Information Technology for Life at Nakhon Pathom Rajabhat University, Thailand
Authors: Sirisawad, Somporn
Abstract: The purpose of this study was to determine whether computer assisted instruction (CAI) in
the Information Technology for Life course taken by first year students at Nakhon Pathom
Rajabhat University (NPRU), Thailand, could be used to teach at least as effectively as
traditional methods. Since CAI has been used successfully in developed countries to
supplement or replace traditional methods of instruction, it was thought that CAI may
present a solution to the lack of instructors in general education courses across the 41
Rajabhat Universities in Thailand. CAI could also facilitate student centred learning, a
key goal of the National Education Act (1999).
One hundred and twenty four incoming freshman students enrolled at NPRU for the
2004 academic year participated in a study comparing the two methods of instruction using
three topics of the Information Technology for Life course. The research questions
examined were (1) are there differences between the groups on the achievement factors
related to CAI usage? and (2) are there differences between the groups on attitude factors
related to CAI and traditional teaching? CAI lessons were developed for the experimental
group as interactive multimedia modules loaded from a CD-ROM; the control group
received traditional lecture instruction.
Pre-test and post-test scores indicated greater learning gains in the CAI group. Comparison
of weak, average and strong students between the two groups showed no difference
in learning outcomes for the weak students, but average and strong students in
the CAI group did better than those of the control group. The results also indicated that
CAI students' retention of content was better than that of students following traditional
learning. There was no significant difference in students' attitudes toward their method
of teaching. Students of both groups felt that overall their method of teaching was very
good. No relationship was found between student performance and their attitude toward
CAI.
Description: University of Technology, Sydney. Faculty of Science.2006-01-01T00:00:00Z