Predicting

[electromania]

I need to write a program that can predict the stock market price indicies. The predictions are based on a very naive method that assumes that the market index is a linear function of time. Real life estimates are, of course, much more complicated and keep many people busy all over the globe. Our assumption that the market index, y, is a linear function of time, x, which means that the index can be written as:
y = ax + b
where a is the slope of the linear function and b is the intersection with the y-axis. Such an assumption is usually referred to as a model, i.e. we use a linear model to predict the stock market indicies.

How do I go about starting this program as I'm not too strong on the maths side.

[Meetee]

I need to write a program that can predict the stock market price indicies. The predictions are based on a very naive method that assumes that the market index is a linear function of time. Real life estimates are, of course, much more complicated and keep many people busy all over the globe. Our assumption that the market index, y, is a linear function of time, x, which means that the index can be written as:
y = ax + b
where a is the slope of the linear function and b is the intersection with the y-axis. Such an assumption is usually referred to as a model, i.e. we use a linear model to predict the stock market indicies.

How do I go about starting this program as I'm not too strong on the maths side.

I think your problem need some more analysis bacause AFAIK this site helps you in technical queries. Can you explain your problem in more detail as this seems to be a problem statement only.

Thanks

[JosAH]

I need to write a program that can predict the stock market price indicies. The predictions are based on a very naive method that assumes that the market index is a linear function of time. Real life estimates are, of course, much more complicated and keep many people busy all over the globe. Our assumption that the market index, y, is a linear function of time, x, which means that the index can be written as:
y = ax + b
where a is the slope of the linear function and b is the intersection with the y-axis. Such an assumption is usually referred to as a model, i.e. we use a linear model to predict the stock market indicies.

How do I go about starting this program as I'm not too strong on the maths side.

The best you can do with this simple assumption (y being a linear function of x)
is to do a linear regression given the data points (x_i, y_i) in order to find a and b.

When you've found those values calculate the correllation coefficient and see that
it is extremely small (very much near zero) telling you that your linear model is
a nono for the data points.

kind regards,

Jos

[electromania]

The best you can do with this simple assumption (y being a linear function of x)
is to do a linear regression given the data points (x_i, y_i) in order to find a and b.

When you've found those values calculate the correllation coefficient and see that
it is extremely small (very much near zero) telling you that your linear model is
a nono for the data points.

kind regards,

Jos

Yes, your right but I need to find out the following tasks

1. Test whether the linear model is a satisfactory description of the market index.
2. If so, what the index will be in the future, i.e. can investments be recommended.

I'll be taking inputs such as x/years and y/index.

I have the formula for the graph and i know how to do it on paper. But I'm not sure where to start in writing a program to do that.

I'm not sure how do i go about calculating this and see if its going towards negative or positive?

[JosAH]

Yes, your right but I need to find out the following tasks

1. Test whether the linear model is a satisfactory description of the market index.
2. If so, what the index will be in the future, i.e. can investments be recommended.

I'll be taking inputs such as x/years and y/index.

I have the formula for the graph and i know how to do it on paper. But I'm not sure where to start in writing a program to do that.

I'm not sure how do i go about calculating this and see if its going towards negative or positive?

Here's the Least Squares algorithm. All you have to do is translate it to C or C++.

kind regards,

Jos