# Identifying short-term highs and lows on time series data

1. ## Identifying short-term highs and lows on time series data

Dear all,

I am examining the patterns on short-term highs and lows on financial time series data. May I know I can identify those short-term highs and lows on a data metric with daily_high and daily_low with excel (below table). Thanks!!!

ST_high ST_low

Date Daily_High Daily_Low
2018/11/19 26,397 26,146
2018/11/20 26,132 25,776
2018/11/21 25,998 25,495
2018/11/22 26,139 25,873
2018/11/23 25,983 25,807
2018/11/26 26,414 26,032
2018/11/27 26,418 26,159
2018/11/28 26,705 26,333
2018/11/29 26,923 26,400
2018/11/30 26,683 26,397
2018/12/03 27,259 27,007
2018/12/04 27,260 26,985
2018/12/05 26,915 26,746
2018/12/06 26,239 26,013
2018/12/07 26,281 26,056
2018/12/10 25,845 25,571
2018/12/11 25,862 25,623
2018/12/12 26,239 26,013
2018/12/13 26,582 26,315
2018/12/14 26,219 26,061
2018/12/17 26,186 26,002
2018/12/18 26,153 25,714
2018/12/19 25,926 25,784
2018/12/20 25,860 25,417
2018/12/21 25,806 25,314
2018/12/24 25,609 25,421

2. ## Re: Identifying short-term highs and lows on time series data

I often find that the most difficult part of this kind of "signal processing" is trying to clearly state in my own mind exactly what I mean by "short term high". It isn't clear to me exactly how you are identifying those 'highs' you have identified. I note that you have chosen not to highlight the small "peaks" on 22 Nov and 7 Dec., so there must be more to your criteria than just "higher than the day before and after".

The overall process I would expect to follow:
1) Identify my criteria for "short term high".
2) Add helper column(s) to identify those criteria.
3) A final column to combine all the criteria together into a single column that identifies the short term highs
4) Conditional formatting to highlight the identified rows.

3. ## Re: Identifying short-term highs and lows on time series data

True that I haven't specified too clearly on my definition on high and low.

The listed table was just my own findings based on graphical presentation, which didn't have clear criteria.

Before I simply tried finding out the local max and min by comparing the highs (lows) at +/- 1 day as well as the max (min) +/- 3 days. But the results aren't so meaningful, because it may create two or more successive max (min), which fail to meet the purpose of identifying the short-term trends.

But it may also be just how financial data behave, as they are often fractal and not following simple arithmetic patterns.

Let me think once again how to define the criteria, as I regard there may be a need to compute the last local max and min at each new rows, or it may even need to correct the last local max and min when new data override the last values.

Again, thanks a lot!

There are currently 1 users browsing this thread. (0 members and 1 guests)

#### Posting Permissions

• You may not post new threads
• You may not post replies
• You may not post attachments
• You may not edit your posts

Search Engine Friendly URLs by vBSEO 3.6.0 RC 1